[lld] [llvm] [clang] [SHT_LLVM_BB_ADDR_MAP] Allow basic-block-sections and labels be used together by decoupling the handling of the two features. (PR #74128)
Greg Clayton via cfe-commits
cfe-commits at lists.llvm.org
Thu Feb 1 13:58:41 PST 2024
https://github.com/clayborg updated https://github.com/llvm/llvm-project/pull/74128
>From 7a7d5481ad5c925d4f31bee3ab66bd1d7d514b73 Mon Sep 17 00:00:00 2001
From: michaelrj-google <71531609+michaelrj-google at users.noreply.github.com>
Date: Thu, 1 Feb 2024 12:54:01 -0800
Subject: [PATCH 01/12] [libc] add bazel support for most of unistd (#80078)
Much of unistd involves modifying files. The tests for these functions
need to use libc_make_test_file_path which didn't exist when they were
first implemented. This patch adds most of unistd to the bazel along
with the corresponding tests. Tests that modify directories had to be
disabled since bazel doesn't seem to handle them properly.
---
libc/test/src/stdio/remove_test.cpp | 9 +-
libc/test/src/sys/stat/mkdirat_test.cpp | 3 +-
libc/test/src/unistd/CMakeLists.txt | 1 +
libc/test/src/unistd/access_test.cpp | 4 +-
libc/test/src/unistd/chdir_test.cpp | 9 +-
libc/test/src/unistd/dup2_test.cpp | 3 +-
libc/test/src/unistd/dup3_test.cpp | 3 +-
libc/test/src/unistd/dup_test.cpp | 3 +-
libc/test/src/unistd/fchdir_test.cpp | 9 +-
libc/test/src/unistd/ftruncate_test.cpp | 5 +-
libc/test/src/unistd/isatty_test.cpp | 6 +-
libc/test/src/unistd/link_test.cpp | 11 +-
libc/test/src/unistd/linkat_test.cpp | 15 +-
libc/test/src/unistd/lseek_test.cpp | 6 +-
libc/test/src/unistd/pread_pwrite_test.cpp | 3 +-
libc/test/src/unistd/read_write_test.cpp | 7 +-
libc/test/src/unistd/readlink_test.cpp | 6 +-
libc/test/src/unistd/readlinkat_test.cpp | 6 +-
libc/test/src/unistd/rmdir_test.cpp | 6 +-
libc/test/src/unistd/symlink_test.cpp | 9 +-
libc/test/src/unistd/symlinkat_test.cpp | 15 +-
libc/test/src/unistd/truncate_test.cpp | 3 +-
libc/test/src/unistd/unlink_test.cpp | 6 +-
libc/test/src/unistd/unlinkat_test.cpp | 9 +-
.../llvm-project-overlay/libc/BUILD.bazel | 254 ++++++++++++-
.../libc/test/src/stdio/BUILD.bazel | 12 +
.../libc/test/src/unistd/BUILD.bazel | 340 ++++++++++++++++++
27 files changed, 707 insertions(+), 56 deletions(-)
create mode 100644 utils/bazel/llvm-project-overlay/libc/test/src/unistd/BUILD.bazel
diff --git a/libc/test/src/stdio/remove_test.cpp b/libc/test/src/stdio/remove_test.cpp
index 67bff906ef2a3..539c23d0fb906 100644
--- a/libc/test/src/stdio/remove_test.cpp
+++ b/libc/test/src/stdio/remove_test.cpp
@@ -23,7 +23,9 @@ TEST(LlvmLibcRemoveTest, CreateAndRemoveFile) {
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/remove.test.file";
+
+ constexpr const char *FILENAME = "remove.test.file";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
@@ -40,7 +42,8 @@ TEST(LlvmLibcRemoveTest, CreateAndRemoveDir) {
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata/remove.test.dir";
+ constexpr const char *FILENAME = "remove.test.dir";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
ASSERT_THAT(LIBC_NAMESPACE::mkdirat(AT_FDCWD, TEST_DIR, S_IRWXU),
Succeeds(0));
@@ -51,5 +54,5 @@ TEST(LlvmLibcRemoveTest, CreateAndRemoveDir) {
TEST(LlvmLibcRemoveTest, RemoveNonExistent) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
- ASSERT_THAT(LIBC_NAMESPACE::remove("testdata/non-existent"), Fails(ENOENT));
+ ASSERT_THAT(LIBC_NAMESPACE::remove("non-existent"), Fails(ENOENT));
}
diff --git a/libc/test/src/sys/stat/mkdirat_test.cpp b/libc/test/src/sys/stat/mkdirat_test.cpp
index ae11af4f6b9b1..cbacc16b402d7 100644
--- a/libc/test/src/sys/stat/mkdirat_test.cpp
+++ b/libc/test/src/sys/stat/mkdirat_test.cpp
@@ -15,7 +15,8 @@
TEST(LlvmLibcMkdiratTest, CreateAndRemove) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata/mkdirat.testdir";
+ constexpr const char *FILENAME = "testdata/mkdirat.testdir";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
ASSERT_THAT(LIBC_NAMESPACE::mkdirat(AT_FDCWD, TEST_DIR, S_IRWXU),
Succeeds(0));
ASSERT_THAT(LIBC_NAMESPACE::rmdir(TEST_DIR), Succeeds(0));
diff --git a/libc/test/src/unistd/CMakeLists.txt b/libc/test/src/unistd/CMakeLists.txt
index 8b9a8db374dd4..3a7fe6f45c091 100644
--- a/libc/test/src/unistd/CMakeLists.txt
+++ b/libc/test/src/unistd/CMakeLists.txt
@@ -159,6 +159,7 @@ add_libc_unittest(
libc.src.unistd.fsync
libc.src.unistd.read
libc.src.unistd.write
+ libc.src.stdio.remove
libc.test.UnitTest.ErrnoSetterMatcher
)
diff --git a/libc/test/src/unistd/access_test.cpp b/libc/test/src/unistd/access_test.cpp
index 6718005846026..1371afa1a00fa 100644
--- a/libc/test/src/unistd/access_test.cpp
+++ b/libc/test/src/unistd/access_test.cpp
@@ -12,6 +12,7 @@
#include "src/unistd/close.h"
#include "src/unistd/unlink.h"
#include "test/UnitTest/ErrnoSetterMatcher.h"
+#include "test/UnitTest/LibcTest.h"
#include "test/UnitTest/Test.h"
#include <sys/stat.h>
@@ -22,7 +23,8 @@ TEST(LlvmLibcAccessTest, CreateAndTest) {
// test that it is accessable in those modes but not in others.
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/access.test";
+ constexpr const char *FILENAME = "access.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/chdir_test.cpp b/libc/test/src/unistd/chdir_test.cpp
index 6676b71eed732..f187a771a1220 100644
--- a/libc/test/src/unistd/chdir_test.cpp
+++ b/libc/test/src/unistd/chdir_test.cpp
@@ -21,9 +21,12 @@ TEST(LlvmLibcChdirTest, ChangeAndOpen) {
// directory and open the same file to make sure that the "chdir" operation
// succeeded.
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata";
- constexpr const char *TEST_FILE = "testdata/chdir.test";
- constexpr const char *TEST_FILE_BASE = "chdir.test";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "testdata/chdir.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
+ constexpr const char *FILENAME3 = "chdir.test";
+ auto TEST_FILE_BASE = libc_make_test_file_path(FILENAME3);
libc_errno = 0;
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_PATH);
diff --git a/libc/test/src/unistd/dup2_test.cpp b/libc/test/src/unistd/dup2_test.cpp
index 70ea1a72bc8e4..6fa09670e90a9 100644
--- a/libc/test/src/unistd/dup2_test.cpp
+++ b/libc/test/src/unistd/dup2_test.cpp
@@ -22,7 +22,8 @@ TEST(LlvmLibcdupTest, ReadAndWriteViaDup) {
constexpr int DUPFD = 0xD0;
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/dup2.test";
+ constexpr const char *FILENAME = "dup2.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/dup3_test.cpp b/libc/test/src/unistd/dup3_test.cpp
index 9c13de47f0be4..cb0761ad0d87d 100644
--- a/libc/test/src/unistd/dup3_test.cpp
+++ b/libc/test/src/unistd/dup3_test.cpp
@@ -28,7 +28,8 @@ TEST(LlvmLibcdupTest, ReadAndWriteViaDup) {
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/dup3.test";
+ constexpr const char *FILENAME = "dup3.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/dup_test.cpp b/libc/test/src/unistd/dup_test.cpp
index 23abf5969fc28..f303cda5caacf 100644
--- a/libc/test/src/unistd/dup_test.cpp
+++ b/libc/test/src/unistd/dup_test.cpp
@@ -21,7 +21,8 @@
TEST(LlvmLibcdupTest, ReadAndWriteViaDup) {
libc_errno = 0;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/dup.test";
+ constexpr const char *FILENAME = "dup.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/fchdir_test.cpp b/libc/test/src/unistd/fchdir_test.cpp
index 0d870e35f51c5..0d1fc810a9e30 100644
--- a/libc/test/src/unistd/fchdir_test.cpp
+++ b/libc/test/src/unistd/fchdir_test.cpp
@@ -21,9 +21,12 @@ TEST(LlvmLibcChdirTest, ChangeAndOpen) {
// directory and open the same file to make sure that the "fchdir" operation
// succeeded.
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata";
- constexpr const char *TEST_FILE = "testdata/fchdir.test";
- constexpr const char *TEST_FILE_BASE = "fchdir.test";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "testdata/fchdir.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
+ constexpr const char *FILENAME3 = "fchdir.test";
+ auto TEST_FILE_BASE = libc_make_test_file_path(FILENAME3);
libc_errno = 0;
int dir_fd = LIBC_NAMESPACE::open(TEST_DIR, O_DIRECTORY);
diff --git a/libc/test/src/unistd/ftruncate_test.cpp b/libc/test/src/unistd/ftruncate_test.cpp
index 50a3508e0a872..d338ceb9ec42f 100644
--- a/libc/test/src/unistd/ftruncate_test.cpp
+++ b/libc/test/src/unistd/ftruncate_test.cpp
@@ -23,7 +23,8 @@ namespace cpp = LIBC_NAMESPACE::cpp;
TEST(LlvmLibcFtruncateTest, CreateAndTruncate) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char TEST_FILE[] = "testdata/ftruncate.test";
+ constexpr const char *FILENAME = "ftruncate.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
constexpr const char WRITE_DATA[] = "hello, ftruncate";
constexpr size_t WRITE_SIZE = sizeof(WRITE_DATA);
char buf[WRITE_SIZE];
@@ -68,5 +69,5 @@ TEST(LlvmLibcFtruncateTest, CreateAndTruncate) {
TEST(LlvmLibcFtruncateTest, TruncateBadFD) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
- ASSERT_THAT(LIBC_NAMESPACE::ftruncate(1, off_t(1)), Fails(EINVAL));
+ ASSERT_THAT(LIBC_NAMESPACE::ftruncate(0, off_t(1)), Fails(EINVAL));
}
diff --git a/libc/test/src/unistd/isatty_test.cpp b/libc/test/src/unistd/isatty_test.cpp
index c0e14c344b436..8f9c58f56b075 100644
--- a/libc/test/src/unistd/isatty_test.cpp
+++ b/libc/test/src/unistd/isatty_test.cpp
@@ -39,7 +39,8 @@ TEST(LlvmLibcIsATTYTest, BadFdTest) {
}
TEST(LlvmLibcIsATTYTest, DevTTYTest) {
- constexpr const char *TTY_FILE = "/dev/tty";
+ constexpr const char *FILENAME = "/dev/tty";
+ auto TTY_FILE = libc_make_test_file_path(FILENAME);
libc_errno = 0;
int fd = LIBC_NAMESPACE::open(TTY_FILE, O_RDONLY);
if (fd > 0) {
@@ -50,7 +51,8 @@ TEST(LlvmLibcIsATTYTest, DevTTYTest) {
}
TEST(LlvmLibcIsATTYTest, FileTest) {
- constexpr const char *TEST_FILE = "testdata/isatty.test";
+ constexpr const char *FILENAME = "isatty.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
libc_errno = 0;
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
diff --git a/libc/test/src/unistd/link_test.cpp b/libc/test/src/unistd/link_test.cpp
index bf6962b397b2c..fd377dd02114c 100644
--- a/libc/test/src/unistd/link_test.cpp
+++ b/libc/test/src/unistd/link_test.cpp
@@ -18,8 +18,10 @@
TEST(LlvmLibcLinkTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/link.test";
- constexpr const char *TEST_FILE_LINK = "testdata/link.test.link";
+ constexpr const char *FILENAME = "link.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "link.test.link";
+ auto TEST_FILE_LINK = libc_make_test_file_path(FILENAME2);
// The test strategy is as follows:
// 1. Create a normal file
@@ -44,7 +46,6 @@ TEST(LlvmLibcLinkTest, CreateAndUnlink) {
TEST(LlvmLibcLinkTest, LinkToNonExistentFile) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
- ASSERT_THAT(
- LIBC_NAMESPACE::link("testdata/non-existent-file", "testdata/bad-link"),
- Fails(ENOENT));
+ ASSERT_THAT(LIBC_NAMESPACE::link("non-existent-file", "bad-link"),
+ Fails(ENOENT));
}
diff --git a/libc/test/src/unistd/linkat_test.cpp b/libc/test/src/unistd/linkat_test.cpp
index 62fbaacac08e4..50384b1ecafb0 100644
--- a/libc/test/src/unistd/linkat_test.cpp
+++ b/libc/test/src/unistd/linkat_test.cpp
@@ -18,11 +18,16 @@
TEST(LlvmLibcLinkatTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata";
- constexpr const char *TEST_FILE = "linkat.test";
- constexpr const char *TEST_FILE_PATH = "testdata/linkat.test";
- constexpr const char *TEST_FILE_LINK = "linkat.test.link";
- constexpr const char *TEST_FILE_LINK_PATH = "testdata/linkat.test.link";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "linkat.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
+ constexpr const char *FILENAME3 = "testdata/linkat.test";
+ auto TEST_FILE_PATH = libc_make_test_file_path(FILENAME3);
+ constexpr const char *FILENAME4 = "linkat.test.link";
+ auto TEST_FILE_LINK = libc_make_test_file_path(FILENAME4);
+ constexpr const char *FILENAME5 = "testdata/linkat.test.link";
+ auto TEST_FILE_LINK_PATH = libc_make_test_file_path(FILENAME5);
// The test strategy is as follows:
// 1. Create a normal file
diff --git a/libc/test/src/unistd/lseek_test.cpp b/libc/test/src/unistd/lseek_test.cpp
index fd73bccf0cb8f..40c0bf07df313 100644
--- a/libc/test/src/unistd/lseek_test.cpp
+++ b/libc/test/src/unistd/lseek_test.cpp
@@ -18,7 +18,8 @@
TEST(LlvmLibcUniStd, LseekTest) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/lseek.test";
+ constexpr const char *FILENAME = "testdata/lseek.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_RDONLY);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
@@ -52,7 +53,8 @@ TEST(LlvmLibcUniStd, LseekTest) {
TEST(LlvmLibcUniStd, LseekFailsTest) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/lseek.test";
+ constexpr const char *FILENAME = "testdata/lseek.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_RDONLY);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/pread_pwrite_test.cpp b/libc/test/src/unistd/pread_pwrite_test.cpp
index 1d17778e550f8..3c42fcc777a68 100644
--- a/libc/test/src/unistd/pread_pwrite_test.cpp
+++ b/libc/test/src/unistd/pread_pwrite_test.cpp
@@ -32,7 +32,8 @@ TEST(LlvmLibcUniStd, PWriteAndPReadBackTest) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/pread_pwrite.test";
+ constexpr const char *FILENAME = "pread_pwrite.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(fd, 0);
diff --git a/libc/test/src/unistd/read_write_test.cpp b/libc/test/src/unistd/read_write_test.cpp
index 1330573582209..8b6ba427a343a 100644
--- a/libc/test/src/unistd/read_write_test.cpp
+++ b/libc/test/src/unistd/read_write_test.cpp
@@ -8,6 +8,7 @@
#include "src/errno/libc_errno.h"
#include "src/fcntl/open.h"
+#include "src/stdio/remove.h"
#include "src/unistd/close.h"
#include "src/unistd/fsync.h"
#include "src/unistd/read.h"
@@ -19,7 +20,9 @@
TEST(LlvmLibcUniStd, WriteAndReadBackTest) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "__unistd_read_write.test";
+ constexpr const char *FILENAME = "__unistd_read_write.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
+
int write_fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(write_fd, 0);
@@ -39,7 +42,7 @@ TEST(LlvmLibcUniStd, WriteAndReadBackTest) {
EXPECT_STREQ(read_buf, HELLO);
ASSERT_THAT(LIBC_NAMESPACE::close(read_fd), Succeeds(0));
- // TODO: 'remove' the test file after the test.
+ ASSERT_THAT(LIBC_NAMESPACE::remove(TEST_FILE), Succeeds(0));
}
TEST(LlvmLibcUniStd, WriteFails) {
diff --git a/libc/test/src/unistd/readlink_test.cpp b/libc/test/src/unistd/readlink_test.cpp
index 19b339ec2b374..1919938082213 100644
--- a/libc/test/src/unistd/readlink_test.cpp
+++ b/libc/test/src/unistd/readlink_test.cpp
@@ -18,8 +18,10 @@ namespace cpp = LIBC_NAMESPACE::cpp;
TEST(LlvmLibcReadlinkTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char LINK_VAL[] = "readlink_test_value";
- constexpr const char LINK[] = "testdata/readlink.test.link";
+ constexpr const char *FILENAME = "readlink_test_value";
+ auto LINK_VAL = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "readlink.test.link";
+ auto LINK = libc_make_test_file_path(FILENAME2);
libc_errno = 0;
// The test strategy is as follows:
diff --git a/libc/test/src/unistd/readlinkat_test.cpp b/libc/test/src/unistd/readlinkat_test.cpp
index 85cbca084778d..03b35c5718ab5 100644
--- a/libc/test/src/unistd/readlinkat_test.cpp
+++ b/libc/test/src/unistd/readlinkat_test.cpp
@@ -20,8 +20,10 @@ namespace cpp = LIBC_NAMESPACE::cpp;
TEST(LlvmLibcReadlinkatTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char LINK_VAL[] = "readlinkat_test_value";
- constexpr const char LINK[] = "testdata/readlinkat.test.link";
+ constexpr const char *FILENAME = "readlinkat_test_value";
+ auto LINK_VAL = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "readlinkat.test.link";
+ auto LINK = libc_make_test_file_path(FILENAME2);
libc_errno = 0;
// The test strategy is as follows:
diff --git a/libc/test/src/unistd/rmdir_test.cpp b/libc/test/src/unistd/rmdir_test.cpp
index 69228ce98c822..93cb0f3f53c1b 100644
--- a/libc/test/src/unistd/rmdir_test.cpp
+++ b/libc/test/src/unistd/rmdir_test.cpp
@@ -16,13 +16,13 @@
TEST(LlvmLibcRmdirTest, CreateAndRemove) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata/rmdir.testdir";
+ constexpr const char *FILENAME = "rmdir.testdir";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
ASSERT_THAT(LIBC_NAMESPACE::mkdir(TEST_DIR, S_IRWXU), Succeeds(0));
ASSERT_THAT(LIBC_NAMESPACE::rmdir(TEST_DIR), Succeeds(0));
}
TEST(LlvmLibcRmdirTest, RemoveNonExistentDir) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
- ASSERT_THAT(LIBC_NAMESPACE::rmdir("testdata/non-existent-dir"),
- Fails(ENOENT));
+ ASSERT_THAT(LIBC_NAMESPACE::rmdir("non-existent-dir"), Fails(ENOENT));
}
diff --git a/libc/test/src/unistd/symlink_test.cpp b/libc/test/src/unistd/symlink_test.cpp
index e0c5979a60876..2be0431e473eb 100644
--- a/libc/test/src/unistd/symlink_test.cpp
+++ b/libc/test/src/unistd/symlink_test.cpp
@@ -18,9 +18,12 @@
TEST(LlvmLibcSymlinkTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE_BASE = "symlink.test";
- constexpr const char *TEST_FILE = "testdata/symlink.test";
- constexpr const char *TEST_FILE_LINK = "testdata/symlink.test.symlink";
+ constexpr const char *FILENAME = "symlink.test";
+ auto TEST_FILE_BASE = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "symlink.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
+ constexpr const char *FILENAME3 = "symlink.test.symlink";
+ auto TEST_FILE_LINK = libc_make_test_file_path(FILENAME3);
// The test strategy is as follows:
// 1. Create a normal file
diff --git a/libc/test/src/unistd/symlinkat_test.cpp b/libc/test/src/unistd/symlinkat_test.cpp
index 1ab4dc8417d6c..4888574b6cd4c 100644
--- a/libc/test/src/unistd/symlinkat_test.cpp
+++ b/libc/test/src/unistd/symlinkat_test.cpp
@@ -18,11 +18,16 @@
TEST(LlvmLibcSymlinkatTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata";
- constexpr const char *TEST_FILE = "symlinkat.test";
- constexpr const char *TEST_FILE_PATH = "testdata/symlinkat.test";
- constexpr const char *TEST_FILE_LINK = "symlinkat.test.link";
- constexpr const char *TEST_FILE_LINK_PATH = "testdata/symlinkat.test.link";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "symlinkat.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
+ constexpr const char *FILENAME3 = "testdata/symlinkat.test";
+ auto TEST_FILE_PATH = libc_make_test_file_path(FILENAME3);
+ constexpr const char *FILENAME4 = "symlinkat.test.link";
+ auto TEST_FILE_LINK = libc_make_test_file_path(FILENAME4);
+ constexpr const char *FILENAME5 = "testdata/symlinkat.test.link";
+ auto TEST_FILE_LINK_PATH = libc_make_test_file_path(FILENAME5);
// The test strategy is as follows:
// 1. Create a normal file
diff --git a/libc/test/src/unistd/truncate_test.cpp b/libc/test/src/unistd/truncate_test.cpp
index 4d316ed1b16fe..a4cb9323bb865 100644
--- a/libc/test/src/unistd/truncate_test.cpp
+++ b/libc/test/src/unistd/truncate_test.cpp
@@ -23,7 +23,8 @@ namespace cpp = LIBC_NAMESPACE::cpp;
TEST(LlvmLibcTruncateTest, CreateAndTruncate) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char TEST_FILE[] = "testdata/truncate.test";
+ constexpr const char *FILENAME = "truncate.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
constexpr const char WRITE_DATA[] = "hello, truncate";
constexpr size_t WRITE_SIZE = sizeof(WRITE_DATA);
char buf[WRITE_SIZE];
diff --git a/libc/test/src/unistd/unlink_test.cpp b/libc/test/src/unistd/unlink_test.cpp
index c2828df65a180..e1ffaab78147e 100644
--- a/libc/test/src/unistd/unlink_test.cpp
+++ b/libc/test/src/unistd/unlink_test.cpp
@@ -17,7 +17,8 @@
TEST(LlvmLibcUnlinkTest, CreateAndUnlink) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_FILE = "testdata/unlink.test";
+ constexpr const char *FILENAME = "unlink.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME);
int write_fd = LIBC_NAMESPACE::open(TEST_FILE, O_WRONLY | O_CREAT, S_IRWXU);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(write_fd, 0);
@@ -27,6 +28,5 @@ TEST(LlvmLibcUnlinkTest, CreateAndUnlink) {
TEST(LlvmLibcUnlinkTest, UnlinkNonExistentFile) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Fails;
- ASSERT_THAT(LIBC_NAMESPACE::unlink("testdata/non-existent-file"),
- Fails(ENOENT));
+ ASSERT_THAT(LIBC_NAMESPACE::unlink("non-existent-file"), Fails(ENOENT));
}
diff --git a/libc/test/src/unistd/unlinkat_test.cpp b/libc/test/src/unistd/unlinkat_test.cpp
index ac32ed4895d6b..2d64a996e6444 100644
--- a/libc/test/src/unistd/unlinkat_test.cpp
+++ b/libc/test/src/unistd/unlinkat_test.cpp
@@ -18,8 +18,10 @@
TEST(LlvmLibcUnlinkatTest, CreateAndDeleteTest) {
using LIBC_NAMESPACE::testing::ErrnoSetterMatcher::Succeeds;
- constexpr const char *TEST_DIR = "testdata";
- constexpr const char *TEST_FILE = "openat.test";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
+ constexpr const char *FILENAME2 = "openat.test";
+ auto TEST_FILE = libc_make_test_file_path(FILENAME2);
int dir_fd = LIBC_NAMESPACE::open(TEST_DIR, O_DIRECTORY);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(dir_fd, 0);
@@ -33,7 +35,8 @@ TEST(LlvmLibcUnlinkatTest, CreateAndDeleteTest) {
}
TEST(LlvmLibcUnlinkatTest, UnlinkatNonExistentFile) {
- constexpr const char *TEST_DIR = "testdata";
+ constexpr const char *FILENAME = "testdata";
+ auto TEST_DIR = libc_make_test_file_path(FILENAME);
int dir_fd = LIBC_NAMESPACE::open(TEST_DIR, O_DIRECTORY);
ASSERT_ERRNO_SUCCESS();
ASSERT_GT(dir_fd, 0);
diff --git a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
index 3d8c5eb178ec9..14b3c29b5824a 100644
--- a/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
+++ b/utils/bazel/llvm-project-overlay/libc/BUILD.bazel
@@ -2502,7 +2502,53 @@ libc_function(
],
)
-############################### unistd targets ###############################
+################################ fcntl targets #################################
+
+libc_function(
+ name = "open",
+ srcs = ["src/fcntl/linux/open.cpp"],
+ hdrs = ["src/fcntl/open.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "openat",
+ srcs = ["src/fcntl/linux/openat.cpp"],
+ hdrs = ["src/fcntl/openat.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "creat",
+ srcs = ["src/fcntl/linux/creat.cpp"],
+ hdrs = ["src/fcntl/creat.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+################################ unistd targets ################################
+
+libc_function(
+ name = "access",
+ srcs = ["src/unistd/linux/access.cpp"],
+ hdrs = ["src/unistd/access.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
libc_function(
name = "chdir",
@@ -2526,6 +2572,50 @@ libc_function(
],
)
+libc_function(
+ name = "dup",
+ srcs = ["src/unistd/linux/dup.cpp"],
+ hdrs = ["src/unistd/dup.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "dup2",
+ srcs = ["src/unistd/linux/dup2.cpp"],
+ hdrs = ["src/unistd/dup2.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "dup3",
+ srcs = ["src/unistd/linux/dup3.cpp"],
+ hdrs = ["src/unistd/dup3.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "environ",
+ srcs = ["src/unistd/environ.cpp"],
+ hdrs = ["src/unistd/environ.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
libc_function(
name = "fchdir",
srcs = ["src/unistd/linux/fchdir.cpp"],
@@ -2559,6 +2649,86 @@ libc_function(
],
)
+# libc_function(
+# name = "getcwd",
+# srcs = ["src/unistd/linux/getcwd.cpp"],
+# hdrs = ["src/unistd/getcwd.h"],
+# deps = [
+# ":__support_common",
+# ":__support_osutil_syscall",
+# ":errno",
+# ],
+# )
+
+libc_function(
+ name = "geteuid",
+ srcs = ["src/unistd/linux/geteuid.cpp"],
+ hdrs = ["src/unistd/geteuid.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "getpid",
+ srcs = ["src/unistd/linux/getpid.cpp"],
+ hdrs = ["src/unistd/getpid.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "getppid",
+ srcs = ["src/unistd/linux/getppid.cpp"],
+ hdrs = ["src/unistd/getppid.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "getuid",
+ srcs = ["src/unistd/linux/getuid.cpp"],
+ hdrs = ["src/unistd/getuid.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+# libc_function(
+# name = "getopt",
+# srcs = ["src/unistd/getopt.cpp"],
+# hdrs = ["src/unistd/getopt.h"],
+# deps = [
+# ":__support_common",
+# ":__support_cpp_optional",
+# ":__support_cpp_string_view",
+# ":__support_file_file",
+# ":__support_osutil_syscall",
+# ":errno",
+# ],
+# )
+
+libc_function(
+ name = "isatty",
+ srcs = ["src/unistd/linux/isatty.cpp"],
+ hdrs = ["src/unistd/isatty.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
libc_function(
name = "link",
srcs = ["src/unistd/linux/link.cpp"],
@@ -2593,6 +2763,28 @@ libc_function(
],
)
+libc_function(
+ name = "pread",
+ srcs = ["src/unistd/linux/pread.cpp"],
+ hdrs = ["src/unistd/pread.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "pwrite",
+ srcs = ["src/unistd/linux/pwrite.cpp"],
+ hdrs = ["src/unistd/pwrite.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
libc_function(
name = "read",
srcs = ["src/unistd/linux/read.cpp"],
@@ -2660,6 +2852,31 @@ libc_function(
],
)
+#TODO: Enable once fullbuild is added to bazel, since this depends on a macro
+# definition in the public header
+
+# libc_function(
+# name = "syscall",
+# srcs = ["src/unistd/linux/syscall.cpp"],
+# hdrs = ["src/unistd/syscall.h"],
+# deps = [
+# ":__support_common",
+# ":__support_osutil_syscall",
+# ":errno",
+# ],
+# )
+
+libc_function(
+ name = "swab",
+ srcs = ["src/unistd/swab.cpp"],
+ hdrs = ["src/unistd/swab.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
libc_function(
name = "truncate",
srcs = ["src/unistd/linux/truncate.cpp"],
@@ -2944,6 +3161,41 @@ libc_function(
],
)
+libc_function(
+ name = "remove",
+ srcs = ["src/stdio/linux/remove.cpp"],
+ hdrs = ["src/stdio/remove.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+############################### sys/stat targets ###############################
+
+libc_function(
+ name = "mkdir",
+ srcs = ["src/sys/stat/linux/mkdir.cpp"],
+ hdrs = ["src/sys/stat/mkdir.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
+libc_function(
+ name = "mkdirat",
+ srcs = ["src/sys/stat/linux/mkdirat.cpp"],
+ hdrs = ["src/sys/stat/mkdirat.h"],
+ deps = [
+ ":__support_common",
+ ":__support_osutil_syscall",
+ ":errno",
+ ],
+)
+
############################## sys/epoll targets ###############################
libc_function(
diff --git a/utils/bazel/llvm-project-overlay/libc/test/src/stdio/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/test/src/stdio/BUILD.bazel
index 3ec2bdb4dcf45..145ef86380b62 100644
--- a/utils/bazel/llvm-project-overlay/libc/test/src/stdio/BUILD.bazel
+++ b/utils/bazel/llvm-project-overlay/libc/test/src/stdio/BUILD.bazel
@@ -120,3 +120,15 @@ libc_test(
"//libc:vfprintf",
],
)
+
+libc_test(
+ name = "remove_test",
+ srcs = ["remove_test.cpp"],
+ libc_function_deps = [
+ "//libc:remove",
+ "//libc:open",
+ "//libc:mkdirat",
+ "//libc:access",
+ "//libc:close",
+ ],
+)
diff --git a/utils/bazel/llvm-project-overlay/libc/test/src/unistd/BUILD.bazel b/utils/bazel/llvm-project-overlay/libc/test/src/unistd/BUILD.bazel
new file mode 100644
index 0000000000000..4549fa2a193fc
--- /dev/null
+++ b/utils/bazel/llvm-project-overlay/libc/test/src/unistd/BUILD.bazel
@@ -0,0 +1,340 @@
+# This file is licensed under the Apache License v2.0 with LLVM Exceptions.
+# See https://llvm.org/LICENSE.txt for license information.
+# SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+
+# Tests for LLVM libc unistd.h functions.
+
+load("//libc/test:libc_test_rules.bzl", "libc_test")
+
+package(default_visibility = ["//visibility:public"])
+
+licenses(["notice"])
+
+libc_test(
+ name = "access_test",
+ srcs = ["access_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:access",
+ "//libc:close",
+ "//libc:unlink",
+ ],
+)
+
+# libc_test(
+# name = "chdir_test",
+# srcs = ["chdir_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:chdir",
+# "//libc:close",
+# ],
+# )
+
+libc_test(
+ name = "dup_test",
+ srcs = ["dup_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:dup",
+ "//libc:read",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+)
+
+libc_test(
+ name = "dup2_test",
+ srcs = ["dup2_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:dup2",
+ "//libc:read",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+)
+
+libc_test(
+ name = "dup3_test",
+ srcs = ["dup3_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:dup3",
+ "//libc:read",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+)
+
+# libc_test(
+# name = "fchdir_test",
+# srcs = ["fchdir_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:fchdir",
+# "//libc:close",
+# ],
+# )
+
+libc_test(
+ name = "ftruncate_test",
+ srcs = ["ftruncate_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:read",
+ "//libc:ftruncate",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+ deps = [
+ "//libc:__support_cpp_string_view",
+ ],
+)
+
+libc_test(
+ name = "pread_pwrite_test",
+ srcs = ["pread_pwrite_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:fsync",
+ "//libc:pread",
+ "//libc:pwrite",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+)
+
+libc_test(
+ name = "read_write_test",
+ srcs = ["read_write_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:fsync",
+ "//libc:read",
+ "//libc:write",
+ "//libc:remove",
+ ],
+)
+
+libc_test(
+ name = "link_test",
+ srcs = ["link_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:link",
+ "//libc:unlink",
+ ],
+)
+
+# libc_test(
+# name = "linkat_test",
+# srcs = ["linkat_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:close",
+# "//libc:linkat",
+# "//libc:unlink",
+# ],
+# )
+
+# libc_test(
+# name = "lseek_test",
+# srcs = ["lseek_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:close",
+# "//libc:lseek",
+# "//libc:read",
+# ],
+# )
+
+# libc_test(
+# name = "rmdir_test",
+# srcs = ["rmdir_test.cpp"],
+# libc_function_deps = [
+# "//libc:mkdir",
+# "//libc:rmdir",
+# ],
+# )
+
+libc_test(
+ name = "swab_test",
+ srcs = ["swab_test.cpp"],
+ libc_function_deps = [
+ "//libc:swab",
+ ],
+ deps = [
+ "//libc:string_utils",
+ ],
+)
+
+# libc_test(
+# name = "readlink_test",
+# srcs = ["readlink_test.cpp"],
+# libc_function_deps = [
+# "//libc:readlink",
+# "//libc:symlink",
+# "//libc:unlink",
+# ],
+# deps = [
+# "//libc:__support_cpp_string_view",
+# ],
+# )
+
+# libc_test(
+# name = "readlinkat_test",
+# srcs = ["readlinkat_test.cpp"],
+# libc_function_deps = [
+# "//libc:readlinkat",
+# "//libc:symlink",
+# "//libc:unlink",
+# ],
+# deps = [
+# "//libc:__support_cpp_string_view",
+# ],
+# )
+
+libc_test(
+ name = "symlink_test",
+ srcs = ["symlink_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:symlink",
+ "//libc:unlink",
+ ],
+)
+
+# libc_test(
+# name = "symlinkat_test",
+# srcs = ["symlinkat_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:close",
+# "//libc:symlinkat",
+# "//libc:unlink",
+# ],
+# )
+
+libc_test(
+ name = "truncate_test",
+ srcs = ["truncate_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:read",
+ "//libc:truncate",
+ "//libc:unlink",
+ "//libc:write",
+ ],
+ deps = [
+ "//libc:__support_cpp_string_view",
+ ],
+)
+
+libc_test(
+ name = "unlink_test",
+ srcs = ["unlink_test.cpp"],
+ libc_function_deps = [
+ "//libc:open",
+ "//libc:close",
+ "//libc:unlink",
+ ],
+)
+
+# libc_test(
+# name = "unlinkat_test",
+# srcs = ["unlinkat_test.cpp"],
+# libc_function_deps = [
+# "//libc:open",
+# "//libc:openat",
+# "//libc:close",
+# "//libc:unlinkat",
+# ],
+# )
+
+libc_test(
+ name = "getpid_test",
+ srcs = ["getpid_test.cpp"],
+ libc_function_deps = [
+ "//libc:getpid",
+ ],
+)
+
+libc_test(
+ name = "getppid_test",
+ srcs = ["getppid_test.cpp"],
+ libc_function_deps = [
+ "//libc:getppid",
+ ],
+)
+
+libc_test(
+ name = "getuid_test",
+ srcs = ["getuid_test.cpp"],
+ libc_function_deps = [
+ "//libc:getuid",
+ ],
+)
+
+libc_test(
+ name = "isatty_test",
+ srcs = ["isatty_test.cpp"],
+ libc_function_deps = [
+ "//libc:isatty",
+ "//libc:open",
+ "//libc:close",
+ ],
+)
+
+libc_test(
+ name = "geteuid_test",
+ srcs = ["geteuid_test.cpp"],
+ libc_function_deps = [
+ "//libc:geteuid",
+ ],
+)
+
+#TODO: Enable once fullbuild is added to bazel, since this depends on a macro
+# definition in the public header
+
+# libc_test(
+# name = "syscall_test",
+# srcs = ["syscall_test.cpp"],
+# libc_function_deps = [
+# "//libc:syscall",
+# ],
+# )
+
+# TODO: add once sysconf is complete
+
+# libc_test(
+# name = "sysconf_test",
+# srcs = ["sysconf_test.cpp"],
+# libc_function_deps = [
+# "//libc:sysconf",
+# ],
+# )
+
+# TODO: add fopencookie and fflush
+
+# libc_test(
+# name = "getopt_test",
+# srcs = ["getopt_test.cpp"],
+# libc_function_deps = [
+# "//libc:getopt",
+# "//libc:fopencookie",
+# "//libc:fflush",
+# ],
+# deps = [
+# "//libc:__support_cpp_array",
+# ],
+# )
>From aaaff74fd1273d88d990ed5f93a714fc0a8ca4be Mon Sep 17 00:00:00 2001
From: Micah Weston <micahsweston at gmail.com>
Date: Thu, 1 Feb 2024 16:02:14 -0500
Subject: [PATCH 02/12] [SHT_LLVM_BB_ADDR_MAP][llvm-readobj] Implements
llvm-readobj handling for PGOAnalysisMap. (#79520)
Adds raw printing of PGOAnalysisMap in llvm-readobj.
I'm leaving the fixme's for a later patch that will provide a 'pretty'
printing for BBFreq and BrProb (i.e. relative frequencies and
probabilities) that will apply to both llvm-readobj and llvm-objdump.
---
.../ELF/bb-addr-map-pgo-analysis-map.test | 214 ++++++++++++++++++
llvm/tools/llvm-readobj/ELFDumper.cpp | 61 +++--
2 files changed, 262 insertions(+), 13 deletions(-)
create mode 100644 llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
diff --git a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
new file mode 100644
index 0000000000000..91b3d7e3902e3
--- /dev/null
+++ b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
@@ -0,0 +1,214 @@
+## This test checks how llvm-readobj prints the PGO Analysis Map with the
+## --bb-addr-map option.
+
+## Check 64-bit:
+# RUN: yaml2obj %s -DBITS=64 -DADDR=0x999999999 -o %t1.x64.o
+# RUN: llvm-readobj %t1.x64.o --bb-addr-map 2>&1 | FileCheck %s -DADDR=0x999999999 -DFILE=%t1.x64.o --check-prefix=CHECK
+# RUN: llvm-readelf %t1.x64.o --bb-addr-map | FileCheck %s --check-prefix=GNU
+
+## Check 32-bit:
+# RUN: yaml2obj %s -DBITS=32 -o %t1.x32.o
+# RUN: llvm-readobj %t1.x32.o --bb-addr-map 2>&1 | FileCheck -DADDR=0x11111 %s -DFILE=%t1.x32.o --check-prefix=CHECK
+# RUN: llvm-readelf %t1.x32.o --bb-addr-map | FileCheck %s --check-prefix=GNU
+
+## Check that a malformed section can be handled.
+# RUN: yaml2obj %s -DBITS=32 -DSIZE=24 -o %t2.o
+# RUN: llvm-readobj %t2.o --bb-addr-map 2>&1 | FileCheck %s -DOFFSET=0x00000018 -DFILE=%t2.o --check-prefix=TRUNCATED
+
+## Check that missing features can be handled.
+# RUN: yaml2obj %s -DBITS=32 -DFEATURE=0x2 -o %t3.o
+# RUN: llvm-readobj %t3.o --bb-addr-map 2>&1 | FileCheck %s -DFILE=%t3.o --check-prefix=INVALIDFT
+
+# CHECK: BBAddrMap [
+# CHECK-NEXT: Function {
+# CHECK-NEXT: At: [[ADDR]]
+# CHECK-NEXT: warning: '[[FILE]]': could not identify function symbol for address ([[ADDR]]) in SHT_LLVM_BB_ADDR_MAP section with index 3
+# CHECK-NEXT: Name: <?>
+# CHECK-NEXT: BB entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 0
+# CHECK-NEXT: Offset: 0x0
+# CHECK-NEXT: Size: 0x1
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: Yes
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 2
+# CHECK-NEXT: Offset: 0x4
+# CHECK-NEXT: Size: 0x4
+# CHECK-NEXT: HasReturn: Yes
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: Yes
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: Yes
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+# CHECK-NEXT: PGO analyses {
+# CHECK-NEXT: FuncEntryCount: 100
+# CHECK-NEXT: PGO BB entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: Frequency: 100
+# CHECK-NEXT: Successors [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 2
+# CHECK-NEXT: Probability: 0xFFFFFFFF
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+# CHECK-NEXT: }
+# CHECK-NEXT: {
+# CHECK-NEXT: Frequency: 100
+# CHECK-NEXT: Successors [
+# CHECK-NEXT: ]
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+# CHECK-NEXT: }
+# CHECK-NEXT: }
+# CHECK-NEXT: Function {
+# CHECK-NEXT: At: 0x22222
+# CHECK-NEXT: Name: foo
+# CHECK-NEXT: BB entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 4
+# CHECK-NEXT: Offset: 0x6
+# CHECK-NEXT: Size: 0x7
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: Yes
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+# CHECK-NEXT: PGO analyses {
+# CHECK-NEXT: FuncEntryCount: 8888
+# CHECK-NEXT: PGO BB entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: Frequency: 9000
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+# CHECK-NEXT: }
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
+
+# GNU: GNUStyle::printBBAddrMaps not implemented
+
+# TRUNCATED: BBAddrMap [
+# TRUNCATED-NEXT: warning: '[[FILE]]': unable to dump SHT_LLVM_BB_ADDR_MAP section with index 3: unable to decode LEB128 at offset [[OFFSET]]: malformed uleb128, extends past end
+# TRUNCATED-NEXT: ]
+## Check that the other valid section is properly dumped.
+# TRUNCATED-NEXT: BBAddrMap [
+# TRUNCATED-NEXT: Function {
+# TRUNCATED-NEXT: At: 0x33333
+# TRUNCATED-NEXT: Name: bar
+# TRUNCATED-NEXT: BB entries [
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 6
+# TRUNCATED-NEXT: Offset: 0x9
+# TRUNCATED-NEXT: Size: 0xA
+# TRUNCATED-NEXT: HasReturn: Yes
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: No
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: Yes
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 7
+# TRUNCATED-NEXT: Offset: 0x1F
+# TRUNCATED-NEXT: Size: 0xD
+# TRUNCATED-NEXT: HasReturn: No
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: Yes
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: No
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: ]
+# TRUNCATED-NEXT: PGO analyses {
+# TRUNCATED-NEXT: FuncEntryCount: 89
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: ]
+
+# INVALIDFT: warning: '[[FILE]]': unable to dump SHT_LLVM_BB_ADDR_MAP section with index 5: unable to decode LEB128 at offset 0x00000010: malformed uleb128, extends past end
+
+--- !ELF
+FileHeader:
+ Class: ELFCLASS[[BITS]]
+ Data: ELFDATA2LSB
+ Type: ET_EXEC
+Sections:
+ - Name: .text
+ Type: SHT_PROGBITS
+ Flags: [SHF_ALLOC]
+ - Name: .text.bar
+ Type: SHT_PROGBITS
+ Flags: [SHF_ALLOC]
+ - Name: .llvm_bb_addr_map
+ Type: SHT_LLVM_BB_ADDR_MAP
+ ShSize: [[SIZE=<none>]]
+ Link: .text
+ Entries:
+ - Version: 2
+ Feature: 0x7
+ Address: [[ADDR=0x11111]]
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
+ - ID: 2
+ AddressOffset: 0x3
+ Size: 0x4
+ Metadata: 0x15
+ - Version: 2
+ Feature: 0x3
+ Address: 0x22222
+ BBEntries:
+ - ID: 4
+ AddressOffset: 0x6
+ Size: 0x7
+ Metadata: 0x8
+ PGOAnalyses:
+ - FuncEntryCount: 100
+ PGOBBEntries:
+ - BBFreq: 100
+ Successors:
+ - ID: 2
+ BrProb: 0xFFFFFFFF
+ - BBFreq: 100
+ Successors: []
+ - FuncEntryCount: 8888
+ PGOBBEntries:
+ - BBFreq: 9000
+ - Name: dummy_section
+ Type: SHT_PROGBITS
+ Size: 16
+ - Name: '.llvm_bb_addr_map (1)'
+ Type: SHT_LLVM_BB_ADDR_MAP
+ Link: .text.bar
+ Entries:
+ - Version: 2
+ Feature: [[FEATURE=0x1]]
+ Address: 0x33333
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0x9
+ Size: 0xa
+ Metadata: 0x1b
+ - ID: 7
+ AddressOffset: 0xc
+ Size: 0xd
+ Metadata: 0xe
+ PGOAnalyses:
+ - FuncEntryCount: 89
+Symbols:
+ - Name: foo
+ Section: .text
+ Type: STT_FUNC
+ Value: 0x22222
+ - Name: bar
+ Section: .text.bar
+ Type: STT_FUNC
+ Value: 0x33333
+
diff --git a/llvm/tools/llvm-readobj/ELFDumper.cpp b/llvm/tools/llvm-readobj/ELFDumper.cpp
index f369a63add114..4bf4640f1a4ca 100644
--- a/llvm/tools/llvm-readobj/ELFDumper.cpp
+++ b/llvm/tools/llvm-readobj/ELFDumper.cpp
@@ -7567,14 +7567,15 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
this->describe(*Sec));
continue;
}
+ std::vector<PGOAnalysisMap> PGOAnalyses;
Expected<std::vector<BBAddrMap>> BBAddrMapOrErr =
- this->Obj.decodeBBAddrMap(*Sec, RelocSec);
+ this->Obj.decodeBBAddrMap(*Sec, RelocSec, &PGOAnalyses);
if (!BBAddrMapOrErr) {
this->reportUniqueWarning("unable to dump " + this->describe(*Sec) +
": " + toString(BBAddrMapOrErr.takeError()));
continue;
}
- for (const BBAddrMap &AM : *BBAddrMapOrErr) {
+ for (const auto &[AM, PAM] : zip_equal(*BBAddrMapOrErr, PGOAnalyses)) {
DictScope D(W, "Function");
W.printHex("At", AM.Addr);
SmallVector<uint32_t> FuncSymIndex =
@@ -7588,17 +7589,51 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
FuncName = this->getStaticSymbolName(FuncSymIndex.front());
W.printString("Name", FuncName);
- ListScope L(W, "BB entries");
- for (const BBAddrMap::BBEntry &BBE : AM.BBEntries) {
- DictScope L(W);
- W.printNumber("ID", BBE.ID);
- W.printHex("Offset", BBE.Offset);
- W.printHex("Size", BBE.Size);
- W.printBoolean("HasReturn", BBE.hasReturn());
- W.printBoolean("HasTailCall", BBE.hasTailCall());
- W.printBoolean("IsEHPad", BBE.isEHPad());
- W.printBoolean("CanFallThrough", BBE.canFallThrough());
- W.printBoolean("HasIndirectBranch", BBE.hasIndirectBranch());
+ {
+ ListScope L(W, "BB entries");
+ for (const BBAddrMap::BBEntry &BBE : AM.BBEntries) {
+ DictScope L(W);
+ W.printNumber("ID", BBE.ID);
+ W.printHex("Offset", BBE.Offset);
+ W.printHex("Size", BBE.Size);
+ W.printBoolean("HasReturn", BBE.hasReturn());
+ W.printBoolean("HasTailCall", BBE.hasTailCall());
+ W.printBoolean("IsEHPad", BBE.isEHPad());
+ W.printBoolean("CanFallThrough", BBE.canFallThrough());
+ W.printBoolean("HasIndirectBranch", BBE.hasIndirectBranch());
+ }
+ }
+
+ if (PAM.FeatEnable.anyEnabled()) {
+ DictScope PD(W, "PGO analyses");
+
+ if (PAM.FeatEnable.FuncEntryCount)
+ W.printNumber("FuncEntryCount", PAM.FuncEntryCount);
+
+ if (PAM.FeatEnable.BBFreq || PAM.FeatEnable.BrProb) {
+ ListScope L(W, "PGO BB entries");
+ for (const PGOAnalysisMap::PGOBBEntry &PBBE : PAM.BBEntries) {
+ DictScope L(W);
+
+ /// FIXME: currently we just emit the raw frequency, it may be
+ /// better to provide an option to scale it by the first entry
+ /// frequence using BlockFrequency::Scaled64 number
+ if (PAM.FeatEnable.BBFreq)
+ W.printNumber("Frequency", PBBE.BlockFreq.getFrequency());
+
+ if (PAM.FeatEnable.BrProb) {
+ ListScope L(W, "Successors");
+ for (const auto &Succ : PBBE.Successors) {
+ DictScope L(W);
+ W.printNumber("ID", Succ.ID);
+ /// FIXME: currently we just emit the raw numerator of the
+ /// probably, it may be better to provide an option to emit it
+ /// as a percentage or other prettied representation
+ W.printHex("Probability", Succ.Prob.getNumerator());
+ }
+ }
+ }
+ }
}
}
}
>From 08fccf80949c649e2c4c81bf9149fd77206002c4 Mon Sep 17 00:00:00 2001
From: Anatoly Trosinenko <atrosinenko at accesssoftek.com>
Date: Fri, 2 Feb 2024 00:02:39 +0300
Subject: [PATCH 03/12] [AArch64][PAC] Expand blend(reg, imm) operation in
aarch64-pauth pass (#74729)
In preparation for implementing code generation for more @llvm.ptrauth.* intrinsics, move the expansion of blend(register, small integer) variant of @llvm.ptrauth.blend to the AArch64PointerAuth pass, where most other PAuth-related code generation takes place.
---
llvm/lib/Target/AArch64/AArch64InstrInfo.td | 7 ++--
.../lib/Target/AArch64/AArch64PointerAuth.cpp | 37 +++++++++++++++++++
.../AArch64/ptrauth-pseudo-instructions.mir | 27 ++++++++++++++
3 files changed, 68 insertions(+), 3 deletions(-)
create mode 100644 llvm/test/CodeGen/AArch64/ptrauth-pseudo-instructions.mir
diff --git a/llvm/lib/Target/AArch64/AArch64InstrInfo.td b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
index 6beacea31cb00..6018d662fe514 100644
--- a/llvm/lib/Target/AArch64/AArch64InstrInfo.td
+++ b/llvm/lib/Target/AArch64/AArch64InstrInfo.td
@@ -1581,6 +1581,9 @@ def PAUTH_PROLOGUE : Pseudo<(outs), (ins), []>, Sched<[]>;
def PAUTH_EPILOGUE : Pseudo<(outs), (ins), []>, Sched<[]>;
}
+def PAUTH_BLEND : Pseudo<(outs GPR64:$disc),
+ (ins GPR64:$addr_disc, i32imm:$int_disc), []>, Sched<[]>;
+
// These pointer authentication instructions require armv8.3a
let Predicates = [HasPAuth] in {
@@ -9317,12 +9320,10 @@ let Predicates = [HasMOPS, HasMTE], Defs = [NZCV], Size = 12, mayLoad = 0, maySt
//-----------------------------------------------------------------------------
// v8.3 Pointer Authentication late patterns
-let Predicates = [HasPAuth] in {
def : Pat<(int_ptrauth_blend GPR64:$Rd, imm64_0_65535:$imm),
- (MOVKXi GPR64:$Rd, (trunc_imm imm64_0_65535:$imm), 48)>;
+ (PAUTH_BLEND GPR64:$Rd, (trunc_imm imm64_0_65535:$imm))>;
def : Pat<(int_ptrauth_blend GPR64:$Rd, GPR64:$Rn),
(BFMXri GPR64:$Rd, GPR64:$Rn, 16, 15)>;
-}
//-----------------------------------------------------------------------------
diff --git a/llvm/lib/Target/AArch64/AArch64PointerAuth.cpp b/llvm/lib/Target/AArch64/AArch64PointerAuth.cpp
index 7509afaeb5fef..d8a2853f68e02 100644
--- a/llvm/lib/Target/AArch64/AArch64PointerAuth.cpp
+++ b/llvm/lib/Target/AArch64/AArch64PointerAuth.cpp
@@ -46,6 +46,13 @@ class AArch64PointerAuth : public MachineFunctionPass {
void authenticateLR(MachineFunction &MF,
MachineBasicBlock::iterator MBBI) const;
+ /// Stores blend(AddrDisc, IntDisc) to the Result register.
+ void emitBlend(MachineBasicBlock::iterator MBBI, Register Result,
+ Register AddrDisc, unsigned IntDisc) const;
+
+ /// Expands PAUTH_BLEND pseudo instruction.
+ void expandPAuthBlend(MachineBasicBlock::iterator MBBI) const;
+
bool checkAuthenticatedLR(MachineBasicBlock::iterator TI) const;
};
@@ -359,6 +366,32 @@ bool AArch64PointerAuth::checkAuthenticatedLR(
return true;
}
+void AArch64PointerAuth::emitBlend(MachineBasicBlock::iterator MBBI,
+ Register Result, Register AddrDisc,
+ unsigned IntDisc) const {
+ MachineBasicBlock &MBB = *MBBI->getParent();
+ DebugLoc DL = MBBI->getDebugLoc();
+
+ if (Result != AddrDisc)
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::ORRXrs), Result)
+ .addReg(AArch64::XZR)
+ .addReg(AddrDisc)
+ .addImm(0);
+
+ BuildMI(MBB, MBBI, DL, TII->get(AArch64::MOVKXi), Result)
+ .addReg(Result)
+ .addImm(IntDisc)
+ .addImm(48);
+}
+
+void AArch64PointerAuth::expandPAuthBlend(
+ MachineBasicBlock::iterator MBBI) const {
+ Register ResultReg = MBBI->getOperand(0).getReg();
+ Register AddrDisc = MBBI->getOperand(1).getReg();
+ unsigned IntDisc = MBBI->getOperand(2).getImm();
+ emitBlend(MBBI, ResultReg, AddrDisc, IntDisc);
+}
+
bool AArch64PointerAuth::runOnMachineFunction(MachineFunction &MF) {
const auto *MFnI = MF.getInfo<AArch64FunctionInfo>();
@@ -390,6 +423,7 @@ bool AArch64PointerAuth::runOnMachineFunction(MachineFunction &MF) {
break;
case AArch64::PAUTH_PROLOGUE:
case AArch64::PAUTH_EPILOGUE:
+ case AArch64::PAUTH_BLEND:
assert(!MI.isBundled());
PAuthPseudoInstrs.push_back(MI.getIterator());
break;
@@ -406,6 +440,9 @@ bool AArch64PointerAuth::runOnMachineFunction(MachineFunction &MF) {
authenticateLR(MF, It);
HasAuthenticationInstrs = true;
break;
+ case AArch64::PAUTH_BLEND:
+ expandPAuthBlend(It);
+ break;
default:
llvm_unreachable("Unhandled opcode");
}
diff --git a/llvm/test/CodeGen/AArch64/ptrauth-pseudo-instructions.mir b/llvm/test/CodeGen/AArch64/ptrauth-pseudo-instructions.mir
new file mode 100644
index 0000000000000..d7fe1953deb47
--- /dev/null
+++ b/llvm/test/CodeGen/AArch64/ptrauth-pseudo-instructions.mir
@@ -0,0 +1,27 @@
+# RUN: llc -mtriple=aarch64--- -run-pass=aarch64-ptrauth -verify-machineinstrs %s -o - | FileCheck %s
+
+# Test the corner cases that cannot be reliably tested using LLVM IR as input.
+
+--- |
+ define i64 @blend_untied(i64 %unused, i64 %ptr_arg) {
+ ret i64 0
+ }
+...
+---
+# Check that the input register is copied to the output one, if not tied.
+
+name: blend_untied
+tracksRegLiveness: true
+body: |
+ bb.0:
+ liveins: $lr, $x0, $x1
+ $x0 = PAUTH_BLEND $x1, 42
+ RET undef $lr
+
+# CHECK: liveins: $lr, $x0, $x1
+# CHECK-NEXT: {{^ +$}}
+# CHECK-NEXT: $x0 = ORRXrs $xzr, $x1, 0
+# CHECK-NEXT: $x0 = MOVKXi $x0, 42, 48
+# CHECK-NEXT: RET undef $lr
+
+...
>From fe408eb5845737f8d6bf5eeed0aca8651991e687 Mon Sep 17 00:00:00 2001
From: =?UTF-8?q?Valentin=20Clement=20=28=E3=83=90=E3=83=AC=E3=83=B3?=
=?UTF-8?q?=E3=82=BF=E3=82=A4=E3=83=B3=20=E3=82=AF=E3=83=AC=E3=83=A1?=
=?UTF-8?q?=E3=83=B3=29?= <clementval at gmail.com>
Date: Thu, 1 Feb 2024 13:03:28 -0800
Subject: [PATCH 04/12] [flang][openacc][openmp] Use #0 from hlfir.declare
value when generating bound ops (#80317)
`getDataOperandBaseAddr` retrieve the address of a value when we need to
generate bound operations. When switching to HLFIR, we did not really
handle the fact that this value was then pointing to the result of a
hlfir.declare. Because of that the `#1` value was being used. `#0` value
is carrying the correct information about lowerbounds and should be
used. This patch updates the `getDataOperandBaseAddr` function to use
the correct result value from hlfir.declare.
---
flang/lib/Lower/DirectivesCommon.h | 5 +
flang/test/Lower/OpenACC/acc-bounds.f90 | 22 ++-
.../test/Lower/OpenACC/acc-data-operands.f90 | 12 +-
flang/test/Lower/OpenACC/acc-data.f90 | 78 ++++----
flang/test/Lower/OpenACC/acc-declare.f90 | 42 ++---
flang/test/Lower/OpenACC/acc-enter-data.f90 | 174 +++++++++---------
flang/test/Lower/OpenACC/acc-exit-data.f90 | 36 ++--
flang/test/Lower/OpenACC/acc-host-data.f90 | 8 +-
flang/test/Lower/OpenACC/acc-kernels-loop.f90 | 48 ++---
flang/test/Lower/OpenACC/acc-kernels.f90 | 68 +++----
.../test/Lower/OpenACC/acc-parallel-loop.f90 | 52 +++---
flang/test/Lower/OpenACC/acc-parallel.f90 | 80 ++++----
flang/test/Lower/OpenACC/acc-private.f90 | 61 +++---
flang/test/Lower/OpenACC/acc-reduction.f90 | 86 ++++-----
flang/test/Lower/OpenACC/acc-serial-loop.f90 | 52 +++---
flang/test/Lower/OpenACC/acc-serial.f90 | 74 ++++----
flang/test/Lower/OpenACC/acc-update.f90 | 78 ++++----
flang/test/Lower/OpenMP/array-bounds.f90 | 21 +--
flang/test/Lower/OpenMP/target.f90 | 22 +--
19 files changed, 520 insertions(+), 499 deletions(-)
diff --git a/flang/lib/Lower/DirectivesCommon.h b/flang/lib/Lower/DirectivesCommon.h
index 0920ff80f487b..bd880376517dd 100644
--- a/flang/lib/Lower/DirectivesCommon.h
+++ b/flang/lib/Lower/DirectivesCommon.h
@@ -31,6 +31,7 @@
#include "flang/Lower/Support/Utils.h"
#include "flang/Optimizer/Builder/BoxValue.h"
#include "flang/Optimizer/Builder/FIRBuilder.h"
+#include "flang/Optimizer/Builder/HLFIRTools.h"
#include "flang/Optimizer/Builder/Todo.h"
#include "flang/Optimizer/HLFIR/HLFIROps.h"
#include "flang/Parser/parse-tree.h"
@@ -614,6 +615,10 @@ getDataOperandBaseAddr(Fortran::lower::AbstractConverter &converter,
fir::FirOpBuilder &builder,
Fortran::lower::SymbolRef sym, mlir::Location loc) {
mlir::Value symAddr = converter.getSymbolAddress(sym);
+ if (auto declareOp =
+ mlir::dyn_cast_or_null<hlfir::DeclareOp>(symAddr.getDefiningOp()))
+ symAddr = declareOp.getResults()[0];
+
// TODO: Might need revisiting to handle for non-shared clauses
if (!symAddr) {
if (const auto *details =
diff --git a/flang/test/Lower/OpenACC/acc-bounds.f90 b/flang/test/Lower/OpenACC/acc-bounds.f90
index 9e8e54bc2f7fa..30ce243764b22 100644
--- a/flang/test/Lower/OpenACC/acc-bounds.f90
+++ b/flang/test/Lower/OpenACC/acc-bounds.f90
@@ -85,13 +85,15 @@ subroutine acc_undefined_extent(a)
!$acc kernels present(a)
!$acc end kernels
end subroutine
+
! CHECK-LABEL: func.func @_QMopenacc_boundsPacc_undefined_extent(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<?xf32>> {fir.bindc_name = "a"}) {
! CHECK: %[[DECL_ARG0:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {uniq_name = "_QMopenacc_boundsFacc_undefined_extentEa"} : (!fir.ref<!fir.array<?xf32>>, !fir.shape<1>) -> (!fir.box<!fir.array<?xf32>>, !fir.ref<!fir.array<?xf32>>)
-! CHECK: %[[ONE:.*]] = arith.constant 1 : index
-! CHECK: %[[ZERO:.*]] = arith.constant 0 : index
-! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[ZERO]] : index) extent(%[[ZERO]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-! CHECK: %[[PRESENT:.*]] = acc.present varPtr(%[[DECL_ARG0]]#1 : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a"}
+! CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#0, %c0{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+! CHECK: %[[UB:.*]] = arith.subi %[[DIMS0]]#1, %c1{{.*}} : index
+! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c0{{.*}} : index) upperbound(%[[UB]] : index) extent(%[[DIMS0]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%c1{{.*}} : index) {strideInBytes = true}
+! CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECL_ARG0]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+! CHECK: %[[PRESENT:.*]] = acc.present varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a"}
! CHECK: acc.kernels dataOperands(%[[PRESENT]] : !fir.ref<!fir.array<?xf32>>)
subroutine acc_multi_strides(a)
@@ -104,15 +106,15 @@ subroutine acc_multi_strides(a)
! CHECK-LABEL: func.func @_QMopenacc_boundsPacc_multi_strides(
! CHECK-SAME: %[[ARG0:.*]]: !fir.box<!fir.array<?x?x?xf32>> {fir.bindc_name = "a"})
! CHECK: %[[DECL_ARG0:.*]]:2 = hlfir.declare %[[ARG0]] {uniq_name = "_QMopenacc_boundsFacc_multi_stridesEa"} : (!fir.box<!fir.array<?x?x?xf32>>) -> (!fir.box<!fir.array<?x?x?xf32>>, !fir.box<!fir.array<?x?x?xf32>>)
-! CHECK: %[[BOX_DIMS0:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#1, %c0{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
+! CHECK: %[[BOX_DIMS0:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#0, %c0{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
! CHECK: %[[BOUNDS0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[BOX_DIMS0]]#1 : index) stride(%[[BOX_DIMS0]]#2 : index) startIdx(%{{.*}} : index) {strideInBytes = true}
! CHECK: %[[STRIDE1:.*]] = arith.muli %[[BOX_DIMS0]]#2, %[[BOX_DIMS0]]#1 : index
-! CHECK: %[[BOX_DIMS1:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#1, %c1{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
+! CHECK: %[[BOX_DIMS1:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#0, %c1{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
! CHECK: %[[BOUNDS1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[BOX_DIMS1]]#1 : index) stride(%[[STRIDE1]] : index) startIdx(%{{.*}} : index) {strideInBytes = true}
! CHECK: %[[STRIDE2:.*]] = arith.muli %[[STRIDE1]], %[[BOX_DIMS1]]#1 : index
-! CHECK: %[[BOX_DIMS2:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#1, %c2{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
+! CHECK: %[[BOX_DIMS2:.*]]:3 = fir.box_dims %[[DECL_ARG0]]#0, %c2{{.*}} : (!fir.box<!fir.array<?x?x?xf32>>, index) -> (index, index, index)
! CHECK: %[[BOUNDS2:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[BOX_DIMS2]]#1 : index) stride(%[[STRIDE2]] : index) startIdx(%{{.*}} : index) {strideInBytes = true}
-! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL_ARG0]]#1 : (!fir.box<!fir.array<?x?x?xf32>>) -> !fir.ref<!fir.array<?x?x?xf32>>
+! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL_ARG0]]#0 : (!fir.box<!fir.array<?x?x?xf32>>) -> !fir.ref<!fir.array<?x?x?xf32>>
! CHECK: %[[PRESENT:.*]] = acc.present varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?x?x?xf32>>) bounds(%29, %33, %37) -> !fir.ref<!fir.array<?x?x?xf32>> {name = "a"}
! CHECK: acc.kernels dataOperands(%[[PRESENT]] : !fir.ref<!fir.array<?x?x?xf32>>) {
@@ -125,9 +127,9 @@ subroutine acc_optional_data(a)
! CHECK-LABEL: func.func @_QMopenacc_boundsPacc_optional_data(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>> {fir.bindc_name = "a", fir.optional}) {
! CHECK: %[[ARG0_DECL:.*]]:2 = hlfir.declare %arg0 {fortran_attrs = #fir.var_attrs<optional, pointer>, uniq_name = "_QMopenacc_boundsFacc_optional_dataEa"} : (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>)
-! CHECK: %[[IS_PRESENT:.*]] = fir.is_present %[[ARG0_DECL]]#1 : (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>) -> i1
+! CHECK: %[[IS_PRESENT:.*]] = fir.is_present %[[ARG0_DECL]]#0 : (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>) -> i1
! CHECK: %[[BOX:.*]] = fir.if %[[IS_PRESENT]] -> (!fir.box<!fir.ptr<!fir.array<?xf32>>>) {
-! CHECK: %[[LOAD:.*]] = fir.load %[[ARG0_DECL]]#1 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
+! CHECK: %[[LOAD:.*]] = fir.load %[[ARG0_DECL]]#0 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
! CHECK: fir.result %[[LOAD]] : !fir.box<!fir.ptr<!fir.array<?xf32>>>
! CHECK: } else {
! CHECK: %[[ABSENT:.*]] = fir.absent !fir.box<!fir.ptr<!fir.array<?xf32>>>
diff --git a/flang/test/Lower/OpenACC/acc-data-operands.f90 b/flang/test/Lower/OpenACC/acc-data-operands.f90
index 8177aeb888ab3..5151e8f6a135c 100644
--- a/flang/test/Lower/OpenACC/acc-data-operands.f90
+++ b/flang/test/Lower/OpenACC/acc-data-operands.f90
@@ -26,16 +26,16 @@ subroutine acc_operand_array_section()
! CHECK: %[[LB:.*]] = arith.constant 0 : index
! CHECK: %[[UB:.*]] = arith.constant 49 : index
! CHECK: %[[BOUND_1_50:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_1_50]]) -> !fir.ref<!fir.array<100xf32>> {name = "a(1:50)"}
+! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_1_50]]) -> !fir.ref<!fir.array<100xf32>> {name = "a(1:50)"}
! CHECK: %[[ONE:.*]] = arith.constant 1 : index
! CHECK: %[[LB:.*]] = arith.constant 50 : index
! CHECK: %[[UB:.*]] = arith.constant 99 : index
! CHECK: %[[BOUND_51_100:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-! CHECK: %[[COPYOUT_CREATE:.*]] = acc.create varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_51_100]]) -> !fir.ref<!fir.array<100xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a(51:100)"}
+! CHECK: %[[COPYOUT_CREATE:.*]] = acc.create varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_51_100]]) -> !fir.ref<!fir.array<100xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a(51:100)"}
! CHECK: acc.data dataOperands(%[[COPYIN]], %[[COPYOUT_CREATE]] : !fir.ref<!fir.array<100xf32>>, !fir.ref<!fir.array<100xf32>>) {
! CHECK: acc.terminator
! CHECK: }
-! CHECK: acc.copyout accPtr(%[[COPYOUT_CREATE]] : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_51_100]]) to varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) {name = "a(51:100)"}
+! CHECK: acc.copyout accPtr(%[[COPYOUT_CREATE]] : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND_51_100]]) to varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) {name = "a(51:100)"}
! Testing array sections of a derived-type component
subroutine acc_operand_array_section_component()
@@ -125,7 +125,7 @@ subroutine acc_operand_array_section_allocatable()
! CHECK-LABEL: func.func @_QMacc_data_operandPacc_operand_array_section_allocatable() {
! CHECK: %[[A:.*]] = fir.alloca !fir.box<!fir.heap<!fir.array<?xf32>>> {bindc_name = "a", uniq_name = "_QMacc_data_operandFacc_operand_array_section_allocatableEa"}
! CHECK: %[[DECLA:.*]]:2 = hlfir.declare %[[A]] {fortran_attrs = #fir.var_attrs<allocatable>
-! CHECK: %[[LOAD_BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+! CHECK: %[[LOAD_BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
! CHECK: %[[LOAD_BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
! CHECK: %[[C0:.*]] = arith.constant 0 : index
! CHECK: %[[DIMS0_0:.*]]:3 = fir.box_dims %[[LOAD_BOX_A_1]], %[[C0]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
@@ -141,7 +141,7 @@ subroutine acc_operand_array_section_allocatable()
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS0_2]]#1 : index) stride(%[[DIMS0_1]]#2 : index) startIdx(%[[DIMS0_0]]#0 : index) {strideInBytes = true}
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[LOAD_BOX_A_0]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>) -> !fir.heap<!fir.array<?xf32>>
! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[BOX_ADDR]] : !fir.heap<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.heap<!fir.array<?xf32>> {name = "a(1:50)"}
-! CHECK: %[[LOAD_BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+! CHECK: %[[LOAD_BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
! CHECK: %[[LOAD_BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
! CHECK: %[[C0:.*]] = arith.constant 0 : index
! CHECK: %[[DIMS0_0:.*]]:3 = fir.box_dims %[[LOAD_BOX_A_1]], %[[C0]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
@@ -177,7 +177,7 @@ subroutine acc_operand_array_section_pointer()
! CHECK-LABEL: func.func @_QMacc_data_operandPacc_operand_array_section_pointer() {
! CHECK: %[[P:.*]] = fir.alloca !fir.box<!fir.ptr<!fir.array<?xf32>>> {bindc_name = "p", uniq_name = "_QMacc_data_operandFacc_operand_array_section_pointerEp"}
! CHECK: %[[DECLP:.*]]:2 = hlfir.declare %[[P]] {fortran_attrs = #fir.var_attrs<pointer>
-! CHECK: %[[LOAD_BOX_P_0:.*]] = fir.load %[[DECLP]]#1 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
+! CHECK: %[[LOAD_BOX_P_0:.*]] = fir.load %[[DECLP]]#0 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
! CHECK: %[[LOAD_BOX_P_1:.*]] = fir.load %[[DECLP]]#1 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
! CHECK: %[[C0:.*]] = arith.constant 0 : index
! CHECK: %[[DIMS0_0:.*]]:3 = fir.box_dims %[[LOAD_BOX_P_1]], %[[C0:.*]] : (!fir.box<!fir.ptr<!fir.array<?xf32>>>, index) -> (index, index, index)
diff --git a/flang/test/Lower/OpenACC/acc-data.f90 b/flang/test/Lower/OpenACC/acc-data.f90
index 5b4ab5a65ee6b..f120be272991a 100644
--- a/flang/test/Lower/OpenACC/acc-data.f90
+++ b/flang/test/Lower/OpenACC/acc-data.f90
@@ -22,55 +22,55 @@ subroutine acc_data
!$acc end data
! CHECK: %[[IF1:.*]] = arith.constant true
-! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
! CHECK: acc.data if(%[[IF1]]) dataOperands(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK:acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK:acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
!$acc data copy(a) if(ifCondition)
!$acc end data
-! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
! CHECK: %[[IFCOND:.*]] = fir.load %{{.*}} : !fir.ref<!fir.logical<4>>
! CHECK: %[[IF2:.*]] = fir.convert %[[IFCOND]] : (!fir.logical<4>) -> i1
! CHECK: acc.data if(%[[IF2]]) dataOperands(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK:acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK:acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
!$acc data copy(a, b, c)
!$acc end data
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.data dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK:acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK:acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK:acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc data copy(a) copy(b) copy(c)
!$acc end data
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.data dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK:acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK:acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK:acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK:acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc data copyin(a) copyin(readonly: b, c)
!$acc end data
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
! CHECK: acc.data dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -78,22 +78,22 @@ subroutine acc_data
!$acc data copyout(a) copyout(zero: b) copyout(c)
!$acc end data
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout_zero>, name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout_zero>, name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
! CHECK: acc.data dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK:acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
-! CHECK:acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copyout_zero>, name = "b"}
-! CHECK:acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
+! CHECK:acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
+! CHECK:acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copyout_zero>, name = "b"}
+! CHECK:acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
!$acc data create(a, b) create(zero: c)
!$acc end data
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.data dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -103,17 +103,17 @@ subroutine acc_data
!$acc data create(c) copy(b) create(a)
!$acc end data
-! CHECK:%[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
-! CHECK:%[[COPY_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK:%[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK:%[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK:%[[COPY_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK:%[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
! CHECK: acc.data dataOperands(%[[CREATE_C]], %[[COPY_B]], %[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
!$acc data no_create(a, b) create(zero: c)
!$acc end data
-! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.data dataOperands(%[[NO_CREATE_A]], %[[NO_CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -122,9 +122,9 @@ subroutine acc_data
!$acc data present(a, b, c)
!$acc end data
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.data dataOperands(%[[PRESENT_A]], %[[PRESENT_B]], %[[PRESENT_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -132,8 +132,8 @@ subroutine acc_data
!$acc data deviceptr(b, c)
!$acc end data
-! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.data dataOperands(%[[DEVICEPTR_B]], %[[DEVICEPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
diff --git a/flang/test/Lower/OpenACC/acc-declare.f90 b/flang/test/Lower/OpenACC/acc-declare.f90
index 94fb4125022de..401b654adeb61 100644
--- a/flang/test/Lower/OpenACC/acc-declare.f90
+++ b/flang/test/Lower/OpenACC/acc-declare.f90
@@ -20,12 +20,12 @@ subroutine acc_declare_copy()
! CHECK-DAG: %[[ALLOCA:.*]] = fir.alloca !fir.array<100xi32> {bindc_name = "a", uniq_name = "_QMacc_declareFacc_declare_copyEa"}
! CHECK-DAG: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_copy>, uniq_name = "_QMacc_declareFacc_declare_copyEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%{{.*}} = %{{.*}}) -> (index, i32) {
! CHECK: }
! CHECK: acc.declare_exit token(%[[TOKEN]]) dataOperands(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>)
-! CHECK: acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) to varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) to varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
! CHECK: return
subroutine acc_declare_create()
@@ -42,7 +42,7 @@ subroutine acc_declare_create()
! CHECK-DAG: %[[ALLOCA:.*]] = fir.alloca !fir.array<100xi32> {bindc_name = "a", uniq_name = "_QMacc_declareFacc_declare_createEa"}
! CHECK-DAG: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_create>, uniq_name = "_QMacc_declareFacc_declare_createEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%{{.*}} = %{{.*}}) -> (index, i32) {
! CHECK: }
@@ -64,7 +64,7 @@ subroutine acc_declare_present(a)
! CHECK-DAG: %[[C1:.*]] = arith.constant 1 : index
! CHECK-DAG: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_present>, uniq_name = "_QMacc_declareFacc_declare_presentEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%[[C1]] : index)
-! CHECK: %[[PRESENT:.*]] = acc.present varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[PRESENT:.*]] = acc.present varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: acc.declare_enter dataOperands(%[[PRESENT]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
@@ -83,9 +83,9 @@ subroutine acc_declare_copyin()
! CHECK: %[[B:.*]] = fir.alloca !fir.array<10xi32> {bindc_name = "b", uniq_name = "_QMacc_declareFacc_declare_copyinEb"}
! CHECK: %[[BDECL:.*]]:2 = hlfir.declare %[[B]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_copyin_readonly>, uniq_name = "_QMacc_declareFacc_declare_copyinEb"} : (!fir.ref<!fir.array<10xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<10xi32>>, !fir.ref<!fir.array<10xi32>>)
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[ADECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[ADECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[BDECL]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[BDECL]]#0 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
! CHECK: acc.declare_enter dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<10xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
@@ -101,11 +101,11 @@ subroutine acc_declare_copyout()
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_copyout()
! CHECK: %[[A:.*]] = fir.alloca !fir.array<100xi32> {bindc_name = "a", uniq_name = "_QMacc_declareFacc_declare_copyoutEa"}
! CHECK: %[[ADECL:.*]]:2 = hlfir.declare %[[A]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_copyout>, uniq_name = "_QMacc_declareFacc_declare_copyoutEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
-! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
! CHECK: acc.declare_exit token(%[[TOKEN]]) dataOperands(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>)
-! CHECK: acc.copyout accPtr(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[ADECL]]#1 : !fir.ref<!fir.array<100xi32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[ADECL]]#0 : !fir.ref<!fir.array<100xi32>>) {name = "a"}
! CHECK: return
subroutine acc_declare_deviceptr(a)
@@ -120,7 +120,7 @@ subroutine acc_declare_deviceptr(a)
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_deviceptr(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}) {
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_deviceptr>, uniq_name = "_QMacc_declareFacc_declare_deviceptrEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
-! CHECK: %[[DEVICEPTR:.*]] = acc.deviceptr varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[DEVICEPTR:.*]] = acc.deviceptr varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: acc.declare_enter dataOperands(%[[DEVICEPTR]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
@@ -136,7 +136,7 @@ subroutine acc_declare_link(a)
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_link(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"})
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_declare_link>, uniq_name = "_QMacc_declareFacc_declare_linkEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
-! CHECK: %[[LINK:.*]] = acc.declare_link varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[LINK:.*]] = acc.declare_link varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: acc.declare_enter dataOperands(%[[LINK]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
@@ -152,7 +152,7 @@ subroutine acc_declare_device_resident(a)
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_device_resident(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"})
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_declare_device_resident>, uniq_name = "_QMacc_declareFacc_declare_device_residentEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
-! CHECK: %[[DEVICERES:.*]] = acc.declare_device_resident varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
+! CHECK: %[[DEVICERES:.*]] = acc.declare_device_resident varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[DEVICERES]] : !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:2 = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%arg{{.*}} = %{{.*}}) -> (index, i32)
! CHECK: acc.declare_exit token(%[[TOKEN]]) dataOperands(%[[DEVICERES]] : !fir.ref<!fir.array<100xi32>>)
@@ -167,7 +167,7 @@ subroutine acc_declare_device_resident2()
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_device_resident2()
! CHECK: %[[ALLOCA:.*]] = fir.alloca !fir.array<100xf32> {bindc_name = "dataparam", uniq_name = "_QMacc_declareFacc_declare_device_resident2Edataparam"}
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_declare_device_resident>, uniq_name = "_QMacc_declareFacc_declare_device_resident2Edataparam"} : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xf32>>, !fir.ref<!fir.array<100xf32>>)
-! CHECK: %[[DEVICERES:.*]] = acc.declare_device_resident varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
+! CHECK: %[[DEVICERES:.*]] = acc.declare_device_resident varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[DEVICERES]] : !fir.ref<!fir.array<100xf32>>)
! CHECK: acc.declare_exit token(%[[TOKEN]]) dataOperands(%[[DEVICERES]] : !fir.ref<!fir.array<100xf32>>)
! CHECK: acc.delete accPtr(%[[DEVICERES]] : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) {dataClause = #acc<data_clause acc_declare_device_resident>, name = "dataparam"}
@@ -181,7 +181,7 @@ subroutine acc_declare_link2()
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_link2()
! CHECK: %[[ALLOCA:.*]] = fir.alloca !fir.array<100xf32> {bindc_name = "dataparam", uniq_name = "_QMacc_declareFacc_declare_link2Edataparam"}
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_declare_link>, uniq_name = "_QMacc_declareFacc_declare_link2Edataparam"} : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xf32>>, !fir.ref<!fir.array<100xf32>>)
-! CHECK: %[[LINK:.*]] = acc.declare_link varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
+! CHECK: %[[LINK:.*]] = acc.declare_link varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
! CHECK: acc.declare_enter dataOperands(%[[LINK]] : !fir.ref<!fir.array<100xf32>>)
subroutine acc_declare_deviceptr2()
@@ -193,7 +193,7 @@ subroutine acc_declare_deviceptr2()
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_deviceptr2()
! CHECK: %[[ALLOCA:.*]] = fir.alloca !fir.array<100xf32> {bindc_name = "dataparam", uniq_name = "_QMacc_declareFacc_declare_deviceptr2Edataparam"}
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_deviceptr>, uniq_name = "_QMacc_declareFacc_declare_deviceptr2Edataparam"} : (!fir.ref<!fir.array<100xf32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xf32>>, !fir.ref<!fir.array<100xf32>>)
-! CHECK: %[[DEVICEPTR:.*]] = acc.deviceptr varPtr(%[[DECL]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
+! CHECK: %[[DEVICEPTR:.*]] = acc.deviceptr varPtr(%[[DECL]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "dataparam"}
! CHECK: acc.declare_enter dataOperands(%[[DEVICEPTR]] : !fir.ref<!fir.array<100xf32>>)
function acc_declare_in_func()
@@ -202,7 +202,7 @@ function acc_declare_in_func()
end function
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_in_func() -> f32 {
-! CHECK: %[[DEVICE_RESIDENT:.*]] = acc.declare_device_resident varPtr(%{{.*}}#1 : !fir.ref<!fir.array<1024xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<1024xf32>> {name = "a"}
+! CHECK: %[[DEVICE_RESIDENT:.*]] = acc.declare_device_resident varPtr(%{{.*}}#0 : !fir.ref<!fir.array<1024xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<1024xf32>> {name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[DEVICE_RESIDENT]] : !fir.ref<!fir.array<1024xf32>>)
! CHECK: %[[LOAD:.*]] = fir.load %{{.*}}#1 : !fir.ref<f32>
! CHECK: acc.declare_exit token(%[[TOKEN]]) dataOperands(%[[DEVICE_RESIDENT]] : !fir.ref<!fir.array<1024xf32>>)
@@ -220,7 +220,7 @@ function acc_declare_in_func2(i)
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_in_func2(%arg0: !fir.ref<i32> {fir.bindc_name = "i"}) -> f32 {
! CHECK: %[[ALLOCA_A:.*]] = fir.alloca !fir.array<1024xf32> {bindc_name = "a", uniq_name = "_QMacc_declareFacc_declare_in_func2Ea"}
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ALLOCA_A]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_create>, uniq_name = "_QMacc_declareFacc_declare_in_func2Ea"} : (!fir.ref<!fir.array<1024xf32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<1024xf32>>, !fir.ref<!fir.array<1024xf32>>)
-! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL_A]]#1 : !fir.ref<!fir.array<1024xf32>>) bounds(%7) -> !fir.ref<!fir.array<1024xf32>> {name = "a"}
+! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL_A]]#0 : !fir.ref<!fir.array<1024xf32>>) bounds(%7) -> !fir.ref<!fir.array<1024xf32>> {name = "a"}
! CHECK: %[[TOKEN:.*]] = acc.declare_enter dataOperands(%[[CREATE]] : !fir.ref<!fir.array<1024xf32>>)
! CHECK: cf.br ^bb1
! CHECK: ^bb1:
@@ -291,14 +291,14 @@ subroutine acc_declare_multiple_directive(a, b)
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "b"}) {
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_copy>, uniq_name = "_QMacc_declareFacc_declare_multiple_directiveEa"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
! CHECK: %[[DECL_B:.*]]:2 = hlfir.declare %[[ARG1]](%{{.*}}) {acc.declare = #acc.declare<dataClause = acc_copyout>, uniq_name = "_QMacc_declareFacc_declare_multiple_directiveEb"} : (!fir.ref<!fir.array<100xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
-! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL_A]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL_B]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[DECL_A]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECL_B]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
! CHECK: acc.declare_enter dataOperands(%[[COPYIN]], %[[CREATE]] : !fir.ref<!fir.array<100xi32>>, !fir.ref<!fir.array<100xi32>>)
! CHECK: %{{.*}}:{{.*}} = fir.do_loop %{{.*}} = %{{.*}} to %{{.*}} step %{{.*}} iter_args(%{{.*}} = %{{.*}}) -> (index, i32) {
-! CHECK: acc.copyout accPtr(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[DECL_B]]#1 : !fir.ref<!fir.array<100xi32>>) {name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[DECL_A]]#1 : !fir.ref<!fir.array<100xi32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[DECL_B]]#0 : !fir.ref<!fir.array<100xi32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN]] : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) to varPtr(%[[DECL_A]]#0 : !fir.ref<!fir.array<100xi32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
subroutine acc_declare_array_section(a)
integer :: a(:)
@@ -312,7 +312,7 @@ subroutine acc_declare_array_section(a)
! CHECK-LABEL: func.func @_QMacc_declarePacc_declare_array_section(
! CHECK-SAME: %[[ARG0:.*]]: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "a"}) {
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]] {uniq_name = "_QMacc_declareFacc_declare_array_sectionEa"} : (!fir.box<!fir.array<?xi32>>) -> (!fir.box<!fir.array<?xi32>>, !fir.box<!fir.array<?xi32>>)
-! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL_A]]#1 {acc.declare = #acc.declare<dataClause = acc_copy>} : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
+! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL_A]]#0 {acc.declare = #acc.declare<dataClause = acc_copy>} : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
! CHECK: %[[COPYIN:.*]] = acc.copyin varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<?xi32>> {dataClause = #acc<data_clause acc_copy>, name = "a(1:10)"}
! CHECK: acc.declare_enter dataOperands(%[[COPYIN]] : !fir.ref<!fir.array<?xi32>>)
diff --git a/flang/test/Lower/OpenACC/acc-enter-data.f90 b/flang/test/Lower/OpenACC/acc-enter-data.f90
index e7c4b8b20528d..2cf50c1b62f19 100644
--- a/flang/test/Lower/OpenACC/acc-enter-data.f90
+++ b/flang/test/Lower/OpenACC/acc-enter-data.f90
@@ -27,7 +27,7 @@ subroutine acc_enter_data
!CHECK: %[[LB:.*]] = arith.constant 0 : index
!CHECK: %[[UB:.*]] = arith.subi %[[EXTENT_C10]], %[[ONE]] : index
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
!$acc enter data create(a) if(.true.)
@@ -38,7 +38,7 @@ subroutine acc_enter_data
!CHECK: %[[LB:.*]] = arith.constant 0 : index
!CHECK: %[[UB:.*]] = arith.subi %[[EXTENT_C10]], %[[ONE]] : index
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: [[IF1:%.*]] = arith.constant true
!CHECK: acc.enter_data if([[IF1]]) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
@@ -50,7 +50,7 @@ subroutine acc_enter_data
!CHECK: %[[LB:.*]] = arith.constant 0 : index
!CHECK: %[[UB:.*]] = arith.subi %[[EXTENT_C10]], %[[ONE]] : index
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: [[IFCOND:%.*]] = fir.load %{{.*}} : !fir.ref<!fir.logical<4>>
!CHECK: [[IF2:%.*]] = fir.convert [[IFCOND]] : (!fir.logical<4>) -> i1
!CHECK: acc.enter_data if([[IF2]]) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
@@ -58,35 +58,35 @@ subroutine acc_enter_data
!$acc enter data create(a) create(b) create(c)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
+!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "c", structured = false}
+!CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "c", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>){{$}}
!$acc enter data create(a) create(b) create(zero: c)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
+!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c", structured = false}
+!CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>){{$}}
!$acc enter data copyin(a) create(b) attach(d)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10_{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
-!CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+!CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "b", structured = false}
+!CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
!CHECK: %[[BOX_ADDR_D:.*]] = fir.box_addr %[[BOX_D]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
!CHECK: %[[ATTACH_D:.*]] = acc.attach varPtr(%[[BOX_ADDR_D]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "d", structured = false}
!CHECK: acc.enter_data dataOperands(%[[COPYIN_A]], %[[CREATE_B]], %[[ATTACH_D]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ptr<f32>){{$}}
@@ -94,25 +94,25 @@ subroutine acc_enter_data
!$acc enter data create(a) async
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) attributes {async}
!$acc enter data create(a) wait
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) attributes {wait}
!$acc enter data create(a) async wait
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) attributes {async, wait}
!$acc enter data create(a) async(1)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[ASYNC1:.*]] = arith.constant 1 : i32
!CHECK: acc.enter_data async(%[[ASYNC1]] : i32) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>)
@@ -120,21 +120,21 @@ subroutine acc_enter_data
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[ASYNC2:.*]] = fir.load %{{.*}} : !fir.ref<i32>
!CHECK: acc.enter_data async(%[[ASYNC2]] : i32) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>)
!$acc enter data create(a) wait(1)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[WAIT1:.*]] = arith.constant 1 : i32
!CHECK: acc.enter_data wait(%[[WAIT1]] : i32) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>)
!$acc enter data create(a) wait(queues: 1, 2)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[WAIT2:.*]] = arith.constant 1 : i32
!CHECK: %[[WAIT3:.*]] = arith.constant 2 : i32
!CHECK: acc.enter_data wait(%[[WAIT2]], %[[WAIT3]] : i32, i32) dataOperands(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>)
@@ -142,7 +142,7 @@ subroutine acc_enter_data
!$acc enter data create(a) wait(devnum: 1: queues: 1, 2)
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a", structured = false}
!CHECK: %[[WAIT4:.*]] = arith.constant 1 : i32
!CHECK: %[[WAIT5:.*]] = arith.constant 2 : i32
!CHECK: %[[WAIT6:.*]] = arith.constant 1 : i32
@@ -151,7 +151,7 @@ subroutine acc_enter_data
!$acc enter data copyin(a(1:10,1:5))
!CHECK: %[[BOUND0:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXTENT_C10]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(1:10,1:5)", structured = false}
+!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND0]], %[[BOUND1]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(1:10,1:5)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>)
!$acc enter data copyin(a(1:,1:5))
@@ -162,7 +162,7 @@ subroutine acc_enter_data
!CHECK: %[[LB2:.*]] = arith.constant 0 : index
!CHECK: %[[UB2:.*]] = arith.constant 4 : index
!CHECK: %[[BOUND2:.*]] = acc.bounds lowerbound(%[[LB2]] : index) upperbound(%[[UB2]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%c1{{.*}} : index)
-!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(1:,1:5)", structured = false}
+!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(1:,1:5)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>)
!$acc enter data copyin(a(:10,1:5))
@@ -173,7 +173,7 @@ subroutine acc_enter_data
!CHECK: %[[LB:.*]] = arith.constant 0 : index
!CHECK: %[[UB2:.*]] = arith.constant 4 : index
!CHECK: %[[BOUND2:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB2]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(:10,1:5)", structured = false}
+!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(:10,1:5)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>)
!$acc enter data copyin(a(:,:))
@@ -183,7 +183,7 @@ subroutine acc_enter_data
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
!CHECK: %[[UB:.*]] = arith.subi %c10{{.*}}, %[[ONE]] : index
!CHECK: %[[BOUND2:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%[[ONE]] : index)
-!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(:,:)", structured = false}
+!CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%[[BOUND1]], %[[BOUND2]]) -> !fir.ref<!fir.array<10x10xf32>> {name = "a(:,:)", structured = false}
end subroutine acc_enter_data
subroutine acc_enter_data_dummy(a, b, n, m)
@@ -213,12 +213,14 @@ subroutine acc_enter_data_dummy(a, b, n, m)
!$acc enter data create(a)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%c10{{.*}} : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a", structured = false}
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xf32>>)
!$acc enter data create(b)
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[EXT_B]] : index) stride(%c1{{.*}} : index) startIdx(%{{.*}} : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "b", structured = false}
+!CHECK: %[[DIMS:.*]]:3 = fir.box_dims %[[DECLB]]#0, %c0{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) extent(%[[DIMS]]#1 : index) stride(%[[DIMS]]#2 : index) startIdx(%{{.*}} : index) {strideInBytes = true}
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "b", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
!$acc enter data create(a(5:10))
@@ -226,37 +228,42 @@ subroutine acc_enter_data_dummy(a, b, n, m)
!CHECK: %[[LB1:.*]] = arith.constant 4 : index
!CHECK: %[[UB1:.*]] = arith.constant 9 : index
!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB1]] : index) upperbound(%[[UB1]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%c1{{.*}} : index)
-!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<10xf32>> {name = "a(5:10)", structured = false}
+!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<10xf32>> {name = "a(5:10)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE1]] : !fir.ref<!fir.array<10xf32>>)
!$acc enter data create(b(n:m))
-!CHECK: %[[ONE:.*]] = arith.constant 1 : index
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %c0{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_N:.*]] = fir.load %[[DECLN]]#0 : !fir.ref<i32>
-!CHECK: %[[CONVERT_N:.*]] = fir.convert %[[LOAD_N]] : (i32) -> index
-!CHECK: %[[LB:.*]] = arith.subi %[[CONVERT_N]], %[[N_IDX]] : index
+!CHECK: %[[N_CONV:.*]] = fir.convert %[[LOAD_N]] : (i32) -> index
+!CHECK: %[[LB:.*]] = arith.subi %[[N_CONV]], %[[N_IDX]] : index
!CHECK: %[[LOAD_M:.*]] = fir.load %[[DECLM]]#0 : !fir.ref<i32>
-!CHECK: %[[CONVERT_M:.*]] = fir.convert %[[LOAD_M]] : (i32) -> index
-!CHECK: %[[UB:.*]] = arith.subi %[[CONVERT_M]], %[[N_IDX]] : index
-!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[ONE]] : index) startIdx(%[[N_IDX]] : index)
-!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(n:m)", structured = false}
+!CHECK: %[[M_CONV:.*]] = fir.convert %[[LOAD_M]] : (i32) -> index
+!CHECK: %[[UB:.*]] = arith.subi %[[M_CONV]], %[[N_IDX]] : index
+!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[N_IDX]] : index) {strideInBytes = true}
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(n:m)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE1]] : !fir.ref<!fir.array<?xf32>>)
!$acc enter data create(b(n:))
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %c0_8 : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_N:.*]] = fir.load %[[DECLN]]#0 : !fir.ref<i32>
!CHECK: %[[CONVERT_N:.*]] = fir.convert %[[LOAD_N]] : (i32) -> index
!CHECK: %[[LB:.*]] = arith.subi %[[CONVERT_N]], %[[N_IDX]] : index
-!CHECK: %[[UB:.*]] = arith.subi %[[EXT_B]], %[[ONE]] : index
-!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[ONE]] : index) startIdx(%[[N_IDX]] : index)
-!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(n:)", structured = false}
+!CHECK: %[[UB:.*]] = arith.subi %[[EXT_B]], %c1{{.*}} : index
+!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[N_IDX]] : index) {strideInBytes = true}
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(n:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE1]] : !fir.ref<!fir.array<?xf32>>)
!$acc enter data create(b(:))
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %c0{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[UB:.*]] = arith.subi %[[EXT_B]], %[[ONE]] : index
-!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[ONE]] : index) startIdx(%[[N_IDX]] : index)
-!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(:)", structured = false}
+!CHECK: %[[BOUND1:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%[[EXT_B]] : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[N_IDX]] : index) {strideInBytes = true}
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[CREATE1:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND1]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE1]] : !fir.ref<!fir.array<?xf32>>)
end subroutine
@@ -275,30 +282,31 @@ subroutine acc_enter_data_non_default_lb()
!CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
!$acc enter data create(a(5:9))
-!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[SECTIONLB:.*]] = arith.constant 5 : index
!CHECK: %[[LB:.*]] = arith.subi %[[SECTIONLB]], %[[BASELB]] : index
!CHECK: %[[SECTIONUB:.*]] = arith.constant 9 : index
!CHECK: %[[UB:.*]] = arith.subi %[[SECTIONUB]], %[[BASELB]] : index
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%[[BASELB]] : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(5:9)", structured = false}
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%{{.*}} : index) startIdx(%[[BASELB]] : index)
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<10xi32>>) -> !fir.ref<!fir.array<10xi32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(5:9)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xi32>>)
!$acc enter data create(a(:))
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[UB:.*]] = arith.subi %[[EXTENT_C10]], %[[ONE]] : index
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[BASELB]] : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(:)", structured = false}
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%{{.*}} : index) startIdx(%[[BASELB]] : index)
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<10xi32>>) -> !fir.ref<!fir.array<10xi32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xi32>>)
!$acc enter data create(a(:6))
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
-!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[SECTIONUB:.*]] = arith.constant 6 : index
!CHECK: %[[UB:.*]] = arith.subi %[[SECTIONUB]], %[[BASELB]] : index
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%[[BASELB]] : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(:6)", structured = false}
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%{{.*}} : index) startIdx(%[[BASELB]] : index)
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<10xi32>>) -> !fir.ref<!fir.array<10xi32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(:6)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xi32>>)
!$acc enter data create(a(4:))
@@ -306,16 +314,19 @@ subroutine acc_enter_data_non_default_lb()
!CHECK: %[[SECTIONLB:.*]] = arith.constant 4 : index
!CHECK: %[[LB:.*]] = arith.subi %[[SECTIONLB]], %[[BASELB]] : index
!CHECK: %[[UB:.*]] = arith.subi %[[EXTENT_C10]], %[[ONE]] : index
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%[[ONE]] : index) startIdx(%[[BASELB]] : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(4:)", structured = false}
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[EXTENT_C10]] : index) stride(%{{.*}} : index) startIdx(%[[BASELB]] : index)
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<10xi32>>) -> !fir.ref<!fir.array<10xi32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "a(4:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xi32>>)
!$acc enter data create(b)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %c0{{.*}} : (!fir.box<!fir.array<10xi32>>, index) -> (index, index, index)
!CHECK: %[[LB:.*]] = arith.constant 0 : index
-!CHECK: %[[UB:.*]] = arith.subi %c10{{.*}}, %[[ONE]] : index
-!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%c10{{.*}} : index) stride(%[[ONE]] : index) startIdx(%c11{{.*}} : index)
-!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "b", structured = false}
+!CHECK: %[[UB:.*]] = arith.subi %[[DIMS0]]#1, %[[ONE]] : index
+!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c0{{.*}} : index) upperbound(%[[UB]] : index) extent(%[[DIMS0]]#1 : index) stride(%{{.*}} : index) startIdx(%c11{{.*}} : index) {strideInBytes = true}
+!CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<10xi32>>) -> !fir.ref<!fir.array<10xi32>>
+!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[ADDR]] : !fir.ref<!fir.array<10xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xi32>> {name = "b", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<10xi32>>)
end subroutine
@@ -338,11 +349,11 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!$acc enter data create(a)
!CHECK: %[[C1:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LB:.*]] = arith.constant 0 : index
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS]]#1, %[[C1]] : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS]]#1 : index) stride(%[[DIMS]]#2 : index) startIdx(%[[C1]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -351,14 +362,14 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[C0:.*]] = arith.constant 0 : index
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS1]]#1, %[[ONE]] : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -366,7 +377,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LB:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
@@ -374,7 +385,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS1]]#1, %[[ONE]] : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(2:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -383,12 +394,12 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[UB:.*]] = arith.constant 3 : index
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLA]]#1, %{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(:4)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -396,13 +407,13 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LB:.*]] = arith.constant 5 : index
!CHECK: %[[UB:.*]] = arith.constant 9 : index
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLA]]#1, %{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(6:10)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -410,7 +421,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_N:.*]] = fir.load %[[DECLN]]#0 : !fir.ref<i32>
!CHECK: %[[CONVERT_N:.*]] = fir.convert %[[LOAD_N]] : (i32) -> index
@@ -421,7 +432,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS]]#1, %[[ONE]] : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(n:)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -430,7 +441,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_M:.*]] = fir.load %[[DECLM]]#0 : !fir.ref<i32>
!CHECK: %[[CONVERT_M:.*]] = fir.convert %[[LOAD_M]] : (i32) -> index
@@ -438,7 +449,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLA]]#1, %{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[BASELB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(:m)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -446,7 +457,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLA]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_N:.*]] = fir.load %[[DECLN]]#0 : !fir.ref<i32>
!CHECK: %[[CONVERT_N:.*]] = fir.convert %[[LOAD_N]] : (i32) -> index
@@ -458,7 +469,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLA]]#1, %{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[ONE]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLA]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "a(n:m)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -466,7 +477,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[LOAD_M:.*]] = fir.load %[[DECLM]]#0 : !fir.ref<i32>
!CHECK: %[[CONVERT_M:.*]] = fir.convert %[[LOAD_M]] : (i32) -> index
@@ -474,7 +485,7 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[DECLB]]#1, %{{.*}} : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[ZERO]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[LB_C10_IDX]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLB]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "b(:m)", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -482,12 +493,12 @@ subroutine acc_enter_data_assumed(a, b, n, m)
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[C0:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#1, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[DECLB]]#0, %[[C0]] : (!fir.box<!fir.array<?xf32>>, index) -> (index, index, index)
!CHECK: %[[C0:.*]] = arith.constant 0 : index
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS0]]#1, %[[ONE]] : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[C0]] : index) upperbound(%[[UB]] : index) extent(%[[DIMS0]]#1 : index) stride(%[[DIMS0]]#2 : index) startIdx(%[[LB_C10_IDX]] : index) {strideInBytes = true}
-!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLB]]#1 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
+!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECLB]]#0 : (!fir.box<!fir.array<?xf32>>) -> !fir.ref<!fir.array<?xf32>>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xf32>> {name = "b", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.ref<!fir.array<?xf32>>)
@@ -505,13 +516,12 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(a)
-!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[C0_0:.*]] = arith.constant 0 : index
-
!CHECK: %[[BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[C0_1:.*]] = arith.constant 0 : index
-!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[BOX_A_1]], %[[C0_1]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
-!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[BOX_A_0]], %[[C0_0]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
+!CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[BOX_A_1]], %c0{{.*}} : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
+!CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[BOX_A_0]], %c0{{.*}} : (!fir.box<!fir.heap<!fir.array<?xf32>>>, index) -> (index, index, index)
!CHECK: %[[UB:.*]] = arith.subi %[[DIMS1]]#1, %c1{{.*}} : index
!CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c0{{.*}} : index) upperbound(%[[UB]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS1]]#2 : index) startIdx(%[[DIMS0]]#0 : index) {strideInBytes = true}
!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX_A_0]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>) -> !fir.heap<!fir.array<?xf32>>
@@ -520,7 +530,7 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(a(:))
-!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
@@ -541,7 +551,7 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(a(2:5))
-!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[C0:.*]] = arith.constant 0 : index
@@ -562,7 +572,7 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(a(3:))
-!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[ONE:.*]] = arith.constant 1 : index
!CHECK: %[[BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
@@ -584,7 +594,7 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(a(:7))
-!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+!CHECK: %[[BOX_A_0:.*]] = fir.load %[[DECLA]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
!CHECK: %[[ZERO:.*]] = arith.constant 0 : index
!CHECK: %[[BOX_A_1:.*]] = fir.load %[[DECLA]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
@@ -604,7 +614,7 @@ subroutine acc_enter_data_allocatable()
!$acc enter data create(i)
-!CHECK: %[[BOX_I:.*]] = fir.load %[[DECLI]]#1 : !fir.ref<!fir.box<!fir.heap<i32>>>
+!CHECK: %[[BOX_I:.*]] = fir.load %[[DECLI]]#0 : !fir.ref<!fir.box<!fir.heap<i32>>>
!CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX_I]] : (!fir.box<!fir.heap<i32>>) -> !fir.heap<i32>
!CHECK: %[[CREATE:.*]] = acc.create varPtr(%[[BOX_ADDR]] : !fir.heap<i32>) -> !fir.heap<i32> {name = "i", structured = false}
!CHECK: acc.enter_data dataOperands(%[[CREATE]] : !fir.heap<i32>)
diff --git a/flang/test/Lower/OpenACC/acc-exit-data.f90 b/flang/test/Lower/OpenACC/acc-exit-data.f90
index fb453322df0aa..6600f08d5bcfe 100644
--- a/flang/test/Lower/OpenACC/acc-exit-data.f90
+++ b/flang/test/Lower/OpenACC/acc-exit-data.f90
@@ -18,86 +18,86 @@ subroutine acc_exit_data
!CHECK: %[[DECLD:.*]]:2 = hlfir.declare %[[D]]
!$acc exit data delete(a)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) if(.true.)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[IF1:.*]] = arith.constant true
!CHECK: acc.exit_data if(%[[IF1]]) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) if(ifCondition)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[IFCOND:.*]] = fir.load %{{.*}} : !fir.ref<!fir.logical<4>>
!CHECK: %[[IF2:.*]] = fir.convert %[[IFCOND]] : (!fir.logical<4>) -> i1
!CHECK: acc.exit_data if(%[[IF2]]) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) delete(b) delete(c)
-!CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
-!CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "b", structured = false}
-!CHECK: %[[DEVPTR_C:.*]] = acc.getdeviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "c", structured = false}
+!CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "b", structured = false}
+!CHECK: %[[DEVPTR_C:.*]] = acc.getdeviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "c", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR_A]], %[[DEVPTR_B]], %[[DEVPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>){{$}}
!CHECK: acc.delete accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!CHECK: acc.delete accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "b", structured = false}
!CHECK: acc.delete accPtr(%[[DEVPTR_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "c", structured = false}
!$acc exit data copyout(a) delete(b) detach(d)
-!CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a", structured = false}
-!CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "b", structured = false}
-!CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+!CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a", structured = false}
+!CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "b", structured = false}
+!CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
!CHECK: %[[D_ADDR:.*]] = fir.box_addr %[[BOX_D]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
!CHECK: %[[DEVPTR_D:.*]] = acc.getdeviceptr varPtr(%[[D_ADDR]] : !fir.ptr<f32>) -> !fir.ptr<f32> {dataClause = #acc<data_clause acc_detach>, name = "d", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR_A]], %[[DEVPTR_B]], %[[DEVPTR_D]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ptr<f32>)
-!CHECK: acc.copyout accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+!CHECK: acc.copyout accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!CHECK: acc.delete accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "b", structured = false}
!CHECK: acc.detach accPtr(%[[DEVPTR_D]] : !fir.ptr<f32>) {name = "d", structured = false}
!$acc exit data delete(a) async
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) attributes {async}
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) wait
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) attributes {wait}
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) async wait
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: acc.exit_data dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) attributes {async, wait}
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) async(1)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[ASYNC1:.*]] = arith.constant 1 : i32
!CHECK: acc.exit_data async(%[[ASYNC1]] : i32) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) async(async)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[ASYNC2:.*]] = fir.load %{{.*}} : !fir.ref<i32>
!CHECK: acc.exit_data async(%[[ASYNC2]] : i32) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) wait(1)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[WAIT1:.*]] = arith.constant 1 : i32
!CHECK: acc.exit_data wait(%[[WAIT1]] : i32) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) wait(queues: 1, 2)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[WAIT2:.*]] = arith.constant 1 : i32
!CHECK: %[[WAIT3:.*]] = arith.constant 2 : i32
!CHECK: acc.exit_data wait(%[[WAIT2]], %[[WAIT3]] : i32, i32) dataOperands(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>)
!CHECK: acc.delete accPtr(%[[DEVPTR]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) {name = "a", structured = false}
!$acc exit data delete(a) wait(devnum: 1: queues: 1, 2)
-!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
+!CHECK: %[[DEVPTR:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_delete>, name = "a", structured = false}
!CHECK: %[[WAIT4:.*]] = arith.constant 1 : i32
!CHECK: %[[WAIT5:.*]] = arith.constant 2 : i32
!CHECK: %[[WAIT6:.*]] = arith.constant 1 : i32
diff --git a/flang/test/Lower/OpenACC/acc-host-data.f90 b/flang/test/Lower/OpenACC/acc-host-data.f90
index 269e461aa5600..bcce84dc7dca4 100644
--- a/flang/test/Lower/OpenACC/acc-host-data.f90
+++ b/flang/test/Lower/OpenACC/acc-host-data.f90
@@ -15,14 +15,14 @@ subroutine acc_host_data()
!$acc end host_data
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: acc.host_data dataOperands(%[[DA]] : !fir.ref<!fir.array<10xf32>>)
!$acc host_data use_device(a) if_present
!$acc end host_data
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: acc.host_data dataOperands(%[[DA]] : !fir.ref<!fir.array<10xf32>>) {
! CHECK: } attributes {ifPresent}
@@ -30,7 +30,7 @@ subroutine acc_host_data()
!$acc end host_data
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: %[[LOAD_IFCOND:.*]] = fir.load %[[DECLIFCOND]]#0 : !fir.ref<!fir.logical<4>>
! CHECK: %[[IFCOND_I1:.*]] = fir.convert %[[LOAD_IFCOND]] : (!fir.logical<4>) -> i1
! CHECK: acc.host_data if(%[[IFCOND_I1]]) dataOperands(%[[DA]] : !fir.ref<!fir.array<10xf32>>)
@@ -39,7 +39,7 @@ subroutine acc_host_data()
!$acc end host_data
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%{{.*}} : index) upperbound(%{{.*}} : index) stride(%{{.*}} : index) startIdx(%{{.*}} : index)
-! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DA:.*]] = acc.use_device varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: acc.host_data dataOperands(%[[DA]] : !fir.ref<!fir.array<10xf32>>)
!$acc host_data use_device(a) if(.false.)
diff --git a/flang/test/Lower/OpenACC/acc-kernels-loop.f90 b/flang/test/Lower/OpenACC/acc-kernels-loop.f90
index d2134e8d2337c..24f537ccec224 100644
--- a/flang/test/Lower/OpenACC/acc-kernels-loop.f90
+++ b/flang/test/Lower/OpenACC/acc-kernels-loop.f90
@@ -291,40 +291,40 @@ subroutine acc_kernels_loop
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc kernels loop copy(a) copy(b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc kernels loop copyin(a) copyin(readonly: b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -337,24 +337,24 @@ subroutine acc_kernels_loop
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
! CHECK: acc.kernels dataOperands(%[[CREATE_A]], %[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
!$acc kernels loop create(b) create(zero: a)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
! CHECK: acc.kernels dataOperands(%[[CREATE_B]], %[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -369,8 +369,8 @@ subroutine acc_kernels_loop
a(i) = b(i)
END DO
-! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.kernels dataOperands(%[[NOCREATE_A]], %[[NOCREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -383,8 +383,8 @@ subroutine acc_kernels_loop
a(i) = b(i)
END DO
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.kernels dataOperands(%[[PRESENT_A]], %[[PRESENT_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -397,8 +397,8 @@ subroutine acc_kernels_loop
a(i) = b(i)
END DO
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.kernels dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -411,10 +411,10 @@ subroutine acc_kernels_loop
a(i) = b(i)
END DO
-! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_F:.*]] = fir.box_addr %[[BOX_F]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_F:.*]] = acc.attach varPtr(%[[BOX_ADDR_F]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "f"}
-! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_G:.*]] = fir.box_addr %[[BOX_G]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_G:.*]] = acc.attach varPtr(%[[BOX_ADDR_G]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "g"}
! CHECK: acc.kernels dataOperands(%[[ATTACH_F]], %[[ATTACH_G]] : !fir.ptr<f32>, !fir.ptr<f32>) {
diff --git a/flang/test/Lower/OpenACC/acc-kernels.f90 b/flang/test/Lower/OpenACC/acc-kernels.f90
index 06194edbe1654..ff4f1d3b54591 100644
--- a/flang/test/Lower/OpenACC/acc-kernels.f90
+++ b/flang/test/Lower/OpenACC/acc-kernels.f90
@@ -182,35 +182,35 @@ subroutine acc_kernels
!$acc kernels copy(a, b, c)
!$acc end kernels
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc kernels copy(a) copy(b) copy(c)
!$acc end kernels
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc kernels copyin(a) copyin(readonly: b, c)
!$acc end kernels
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -218,22 +218,22 @@ subroutine acc_kernels
!$acc kernels copyout(a) copyout(zero: b) copyout(c)
!$acc end kernels
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
-! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
!$acc kernels create(a, b) create(zero: c)
!$acc end kernels
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -244,9 +244,9 @@ subroutine acc_kernels
!$acc kernels no_create(a, b) create(zero: c)
!$acc end kernels
-! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.kernels dataOperands(%[[NO_CREATE_A]], %[[NO_CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -254,9 +254,9 @@ subroutine acc_kernels
!$acc kernels present(a, b, c)
!$acc end kernels
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.kernels dataOperands(%[[PRESENT_A]], %[[PRESENT_B]], %[[PRESENT_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -264,8 +264,8 @@ subroutine acc_kernels
!$acc kernels deviceptr(a) deviceptr(c)
!$acc end kernels
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.kernels dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.terminator
! CHECK-NEXT: }{{$}}
@@ -273,10 +273,10 @@ subroutine acc_kernels
!$acc kernels attach(d, e)
!$acc end kernels
-! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_D:.*]] = fir.box_addr %[[BOX_D]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_D:.*]] = acc.attach varPtr(%[[BOX_ADDR_D]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "d"}
-! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_E:.*]] = fir.box_addr %[[BOX_E]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_E:.*]] = acc.attach varPtr(%[[BOX_ADDR_E]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "e"}
! CHECK: acc.kernels dataOperands(%[[ATTACH_D]], %[[ATTACH_E]] : !fir.ptr<f32>, !fir.ptr<f32>) {
diff --git a/flang/test/Lower/OpenACC/acc-parallel-loop.f90 b/flang/test/Lower/OpenACC/acc-parallel-loop.f90
index 24e443a20c895..5968b24782fb4 100644
--- a/flang/test/Lower/OpenACC/acc-parallel-loop.f90
+++ b/flang/test/Lower/OpenACC/acc-parallel-loop.f90
@@ -291,40 +291,40 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc parallel loop copy(a) copy(b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc parallel loop copyin(a) copyin(readonly: b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -337,24 +337,24 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
! CHECK: acc.parallel dataOperands(%[[CREATE_A]], %[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
!$acc parallel loop create(b) create(zero: a)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
! CHECK: acc.parallel dataOperands(%[[CREATE_B]], %[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -369,8 +369,8 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.parallel dataOperands(%[[NOCREATE_A]], %[[NOCREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -383,8 +383,8 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.parallel dataOperands(%[[PRESENT_A]], %[[PRESENT_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -397,8 +397,8 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.parallel dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -411,10 +411,10 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_F:.*]] = fir.box_addr %[[BOX_F]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_F:.*]] = acc.attach varPtr(%[[BOX_ADDR_F]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "f"}
-! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_G:.*]] = fir.box_addr %[[BOX_G]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_G:.*]] = acc.attach varPtr(%[[BOX_ADDR_G]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "g"}
! CHECK: acc.parallel dataOperands(%[[ATTACH_F]], %[[ATTACH_G]] : !fir.ptr<f32>, !fir.ptr<f32>) {
@@ -429,9 +429,9 @@ subroutine acc_parallel_loop
a(i) = b(i)
END DO
-! CHECK: %[[ACC_PRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[ACC_PRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.parallel firstprivate(@firstprivatization_section_ext10_ref_10xf32 -> %[[ACC_PRIVATE_B]] : !fir.ref<!fir.array<10xf32>>) {
-! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: acc.loop private({{.*}}@privatization_ref_10xf32 -> %[[ACC_PRIVATE_A]] : !fir.ref<!fir.array<10xf32>>)
! CHECK-NOT: fir.do_loop
! CHECK: acc.yield
diff --git a/flang/test/Lower/OpenACC/acc-parallel.f90 b/flang/test/Lower/OpenACC/acc-parallel.f90
index 6b37ecb5fab9a..4b18a8d037f22 100644
--- a/flang/test/Lower/OpenACC/acc-parallel.f90
+++ b/flang/test/Lower/OpenACC/acc-parallel.f90
@@ -211,36 +211,36 @@ subroutine acc_parallel
!$acc parallel copy(a, b, c)
!$acc end parallel
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc parallel copy(a) copy(b) copy(c)
!$acc end parallel
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc parallel copyin(a) copyin(readonly: b, c)
!$acc end parallel
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -248,22 +248,22 @@ subroutine acc_parallel
!$acc parallel copyout(a) copyout(zero: b) copyout(c)
!$acc end parallel
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
-! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
!$acc parallel create(a, b) create(zero: c)
!$acc end parallel
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -273,17 +273,17 @@ subroutine acc_parallel
!$acc parallel create(c) copy(b) create(a)
!$acc end parallel
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
-! CHECK: %[[COPY_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[COPY_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
! CHECK: acc.parallel dataOperands(%[[CREATE_C]], %[[COPY_B]], %[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
!$acc parallel no_create(a, b) create(zero: c)
!$acc end parallel
-! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.parallel dataOperands(%[[NO_CREATE_A]], %[[NO_CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -293,9 +293,9 @@ subroutine acc_parallel
!$acc parallel present(a, b, c)
!$acc end parallel
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.parallel dataOperands(%[[PRESENT_A]], %[[PRESENT_B]], %[[PRESENT_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -303,8 +303,8 @@ subroutine acc_parallel
!$acc parallel deviceptr(a) deviceptr(c)
!$acc end parallel
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.parallel dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -312,10 +312,10 @@ subroutine acc_parallel
!$acc parallel attach(d, e)
!$acc end parallel
-! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_D:.*]] = fir.box_addr %[[BOX_D]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_D:.*]] = acc.attach varPtr(%[[BOX_ADDR_D]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "d"}
-! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_E:.*]] = fir.box_addr %[[BOX_E]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_E:.*]] = acc.attach varPtr(%[[BOX_ADDR_E]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "e"}
! CHECK: acc.parallel dataOperands(%[[ATTACH_D]], %[[ATTACH_E]] : !fir.ptr<f32>, !fir.ptr<f32>) {
@@ -327,9 +327,9 @@ subroutine acc_parallel
!$acc parallel private(a) firstprivate(b) private(c)
!$acc end parallel
-! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[ACC_PRIVATE_C:.*]] = acc.private varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[ACC_PRIVATE_C:.*]] = acc.private varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.parallel firstprivate(@firstprivatization_section_ext10xext10_ref_10x10xf32 -> %[[ACC_FPRIVATE_B]] : !fir.ref<!fir.array<10x10xf32>>) private(@privatization_ref_10x10xf32 -> %[[ACC_PRIVATE_A]] : !fir.ref<!fir.array<10x10xf32>>, @privatization_ref_10x10xf32 -> %[[ACC_PRIVATE_C]] : !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
diff --git a/flang/test/Lower/OpenACC/acc-private.f90 b/flang/test/Lower/OpenACC/acc-private.f90
index c0bcb5e2ffd56..29074cbaedc91 100644
--- a/flang/test/Lower/OpenACC/acc-private.f90
+++ b/flang/test/Lower/OpenACC/acc-private.f90
@@ -11,19 +11,19 @@
! CHECK: acc.yield %[[DECLARE]]#0 : !fir.ref<!fir.array<10xf32>>
! CHECK: }
-! CHECK-LABEL: acc.firstprivate.recipe @firstprivatization_ref_UxUx2xi32 : !fir.ref<!fir.array<?x?x2xi32>> init {
-! CHECK: ^bb0(%[[ARG0:.*]]: !fir.ref<!fir.array<?x?x2xi32>>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: index, %[[ARG3:.*]]: index):
-! CHECK: %[[SHAPE:.*]] = fir.shape %[[ARG1]], %[[ARG2]], %[[ARG3]] : (index, index, index) -> !fir.shape<3>
-! CHECK: %[[TEMP:.*]] = fir.alloca !fir.array<?x?x2xi32>, %[[ARG1]], %[[ARG2]], %[[ARG3]]
-! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[TEMP]](%[[SHAPE]]) {uniq_name = "acc.private.init"} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
+! CHECK-LABEL: acc.firstprivate.recipe @firstprivatization_box_UxUx2xi32 : !fir.box<!fir.array<?x?x2xi32>> init {
+! CHECK: ^bb0(%[[ARG0:.*]]: !fir.box<!fir.array<?x?x2xi32>>):
+! CHECK: %[[DIM0:.*]]:3 = fir.box_dims %arg0, %c0{{.*}} : (!fir.box<!fir.array<?x?x2xi32>>, index) -> (index, index, index)
+! CHECK: %[[DIM1:.*]]:3 = fir.box_dims %arg0, %c1{{.*}} : (!fir.box<!fir.array<?x?x2xi32>>, index) -> (index, index, index)
+! CHECK: %[[SHAPE:.*]] = fir.shape %[[DIM0]]#1, %[[DIM1]]#1, %c2{{.*}} : (index, index, index) -> !fir.shape<3>
+! CHECK: %[[TEMP:.*]] = fir.allocmem !fir.array<?x?x2xi32>, %[[DIM0]]#1, %[[DIM1]]#1 {bindc_name = ".tmp", uniq_name = ""}
+! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[TEMP]](%[[SHAPE]]) {uniq_name = ".tmp"} : (!fir.heap<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.heap<!fir.array<?x?x2xi32>>)
! CHECK: acc.yield %[[DECL]]#0 : !fir.box<!fir.array<?x?x2xi32>>
! CHECK: } copy {
-! CHECK: ^bb0(%[[SRC:.*]]: !fir.ref<!fir.array<?x?x2xi32>>, %[[DST:.*]]: !fir.ref<!fir.array<?x?x2xi32>>, %[[LB0:.*]]: index, %[[UB0:.*]]: index, %[[STEP0:.*]]: index, %[[LB1:.*]]: index, %[[UB1:.*]]: index, %[[STEP1:.*]]: index, %[[LB2:.*]]: index, %[[UB2:.*]]: index, %[[STEP2:.*]]: index):
+! CHECK: ^bb0(%[[ARG0:.*]]: !fir.box<!fir.array<?x?x2xi32>>, %[[ARG1:.*]]: !fir.box<!fir.array<?x?x2xi32>>, %[[LB0:.*]]: index, %[[UB0:.*]]: index, %[[STEP0:.*]]: index, %[[LB1:.*]]: index, %[[UB1:.*]]: index, %[[STEP1:.*]]: index, %[[LB2:.*]]: index, %[[UB2:.*]]: index, %[[STEP2:.*]]: index):
! CHECK: %[[SHAPE:.*]] = fir.shape %{{.*}}, %{{.*}}, %{{.*}} : (index, index, index) -> !fir.shape<3>
-! CHECK: %[[DECL_SRC:.*]]:2 = hlfir.declare %[[SRC]](%[[SHAPE]]) {uniq_name = ""} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
-! CHECK: %[[DECL_DST:.*]]:2 = hlfir.declare %[[DST]](%[[SHAPE]]) {uniq_name = ""} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
-! CHECK: %[[DES_SRC:.*]] = hlfir.designate %[[DECL_SRC]]#0 (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]], %[[LB2]]:%[[UB2]]:%[[STEP2]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?x2xi32>>, index, index, index, index, index, index, index, index, index, !fir.shape<3>) -> !fir.box<!fir.array<?x?x2xi32>>
-! CHECK: %[[DES_DST:.*]] = hlfir.designate %[[DECL_DST]]#0 (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]], %[[LB2]]:%[[UB2]]:%[[STEP2]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?x2xi32>>, index, index, index, index, index, index, index, index, index, !fir.shape<3>) -> !fir.box<!fir.array<?x?x2xi32>>
+! CHECK: %[[DES_SRC:.*]] = hlfir.designate %[[ARG0]] (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]], %[[LB2]]:%[[UB2]]:%[[STEP2]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?x2xi32>>, index, index, index, index, index, index, index, index, index, !fir.shape<3>) -> !fir.box<!fir.array<?x?x2xi32>>
+! CHECK: %[[DES_DST:.*]] = hlfir.designate %[[ARG1]] (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]], %[[LB2]]:%[[UB2]]:%[[STEP2]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?x2xi32>>, index, index, index, index, index, index, index, index, index, !fir.shape<3>) -> !fir.box<!fir.array<?x?x2xi32>>
! CHECK: hlfir.assign %[[DES_SRC]] to %[[DES_DST]] : !fir.box<!fir.array<?x?x2xi32>>, !fir.box<!fir.array<?x?x2xi32>>
! CHECK: acc.terminator
! CHECK: }
@@ -66,11 +66,13 @@
! CHECK: acc.terminator
! CHECK: }
-! CHECK-LABEL: acc.private.recipe @privatization_ref_UxUx2xi32 : !fir.ref<!fir.array<?x?x2xi32>> init {
-! CHECK: ^bb0(%[[ARG0:.*]]: !fir.ref<!fir.array<?x?x2xi32>>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: index, %[[ARG3:.*]]: index):
-! CHECK: %[[SHAPE:.*]] = fir.shape %[[ARG1]], %[[ARG2]], %[[ARG3]] : (index, index, index) -> !fir.shape<3>
-! CHECK: %[[ALLOCA:.*]] = fir.alloca !fir.array<?x?x2xi32>, %[[ARG1]], %[[ARG2]], %[[ARG3]]
-! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]](%[[SHAPE]]) {uniq_name = "acc.private.init"} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
+! CHECK-LABEL: acc.private.recipe @privatization_box_UxUx2xi32 : !fir.box<!fir.array<?x?x2xi32>> init {
+! CHECK: ^bb0(%[[ARG0:.*]]: !fir.box<!fir.array<?x?x2xi32>>):
+! CHECK: %[[DIM0:.*]]:3 = fir.box_dims %arg0, %c0{{.*}} : (!fir.box<!fir.array<?x?x2xi32>>, index) -> (index, index, index)
+! CHECK: %[[DIM1:.*]]:3 = fir.box_dims %arg0, %c1{{.*}} : (!fir.box<!fir.array<?x?x2xi32>>, index) -> (index, index, index)
+! CHECK: %[[SHAPE:.*]] = fir.shape %[[DIM0]]#1, %[[DIM1]]#1, %c2{{.*}} : (index, index, index) -> !fir.shape<3>
+! CHECK: %[[TEMP:.*]] = fir.allocmem !fir.array<?x?x2xi32>, %[[DIM0]]#1, %[[DIM1]]#1 {bindc_name = ".tmp", uniq_name = ""}
+! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[TEMP]](%[[SHAPE]]) {uniq_name = ".tmp"} : (!fir.heap<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.heap<!fir.array<?x?x2xi32>>)
! CHECK: acc.yield %[[DECL]]#0 : !fir.box<!fir.array<?x?x2xi32>>
! CHECK: }
@@ -188,7 +190,7 @@ program acc_private
a(i) = b(i) + c
END DO
-! CHECK: %[[C_PRIVATE:.*]] = acc.private varPtr(%[[DECLC]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "c"}
+! CHECK: %[[C_PRIVATE:.*]] = acc.private varPtr(%[[DECLC]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "c"}
! CHECK: acc.loop private({{.*}}@privatization_ref_i32 -> %[[C_PRIVATE]] : !fir.ref<i32>)
! CHECK: acc.yield
@@ -202,7 +204,7 @@ program acc_private
! CHECK: %[[LB:.*]] = arith.constant 0 : index
! CHECK: %[[UB:.*]] = arith.subi %{{.*}}, %[[C1]] : index
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[B_PRIVATE:.*]] = acc.private varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
+! CHECK: %[[B_PRIVATE:.*]] = acc.private varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
! CHECK: acc.loop private({{.*}}@privatization_ref_100xf32 -> %[[B_PRIVATE]] : !fir.ref<!fir.array<100xf32>>)
! CHECK: acc.yield
@@ -216,7 +218,7 @@ program acc_private
! CHECK: %[[LB:.*]] = arith.constant 0 : index
! CHECK: %[[UB:.*]] = arith.constant 49 : index
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[B_PRIVATE:.*]] = acc.private varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<50xf32>> {name = "b(1:50)"}
+! CHECK: %[[B_PRIVATE:.*]] = acc.private varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<50xf32>> {name = "b(1:50)"}
! CHECK: acc.loop private({{.*}}@privatization_ref_50xf32 -> %[[B_PRIVATE]] : !fir.ref<!fir.array<50xf32>>)
!$acc parallel loop firstprivate(c)
@@ -225,7 +227,7 @@ program acc_private
a(i) = b(i) + c
END DO
-! CHECK: %[[FP_C:.*]] = acc.firstprivate varPtr(%[[DECLC]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "c"}
+! CHECK: %[[FP_C:.*]] = acc.firstprivate varPtr(%[[DECLC]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "c"}
! CHECK: acc.parallel firstprivate(@firstprivatization_ref_i32 -> %[[FP_C]] : !fir.ref<i32>)
! CHECK: acc.yield
@@ -239,7 +241,7 @@ program acc_private
! CHECK: %[[LB:.*]] = arith.constant 0 : index
! CHECK: %[[UB:.*]] = arith.subi %{{.*}}, %[[C1]] : index
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[FP_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
+! CHECK: %[[FP_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
! CHECK: acc.parallel firstprivate(@firstprivatization_section_ext100_ref_100xf32 -> %[[FP_B]] : !fir.ref<!fir.array<100xf32>>)
! CHECK: acc.yield
@@ -253,7 +255,7 @@ program acc_private
! CHECK: %[[LB:.*]] = arith.constant 50 : index
! CHECK: %[[UB:.*]] = arith.constant 99 : index
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%{{.*}} : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[FP_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<50xf32>> {name = "b(51:100)"}
+! CHECK: %[[FP_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<50xf32>> {name = "b(51:100)"}
! CHECK: acc.parallel firstprivate(@firstprivatization_section_lb50.ub99_ref_50xf32 -> %[[FP_B]] : !fir.ref<!fir.array<50xf32>>)
end program
@@ -271,7 +273,7 @@ subroutine acc_private_assumed_shape(a, n)
! CHECK-SAME: %[[ARG0:.*]]: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "a"}
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]] {uniq_name = "_QFacc_private_assumed_shapeEa"} : (!fir.box<!fir.array<?xi32>>) -> (!fir.box<!fir.array<?xi32>>, !fir.box<!fir.array<?xi32>>)
! CHECK: acc.parallel {
-! CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECL_A]]#1 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
+! CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECL_A]]#0 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
! CHECK: %[[PRIVATE:.*]] = acc.private varPtr(%[[ADDR]] : !fir.ref<!fir.array<?xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<?xi32>> {name = "a"}
! CHECK: acc.loop private({{.*}}@privatization_box_Uxi32 -> %[[PRIVATE]] : !fir.ref<!fir.array<?xi32>>)
@@ -293,7 +295,7 @@ subroutine acc_private_allocatable_array(a, n)
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>> {fir.bindc_name = "a"}
! CHECK: %[[DECLA_A:.*]]:2 = hlfir.declare %[[ARG0]] {fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFacc_private_allocatable_arrayEa"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>)
! CHECK: acc.parallel {
-! CHECK: %[[BOX:.*]] = fir.load %[[DECLA_A]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>
+! CHECK: %[[BOX:.*]] = fir.load %[[DECLA_A]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xi32>>>>
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX]] : (!fir.box<!fir.heap<!fir.array<?xi32>>>) -> !fir.heap<!fir.array<?xi32>>
! CHECK: %[[PRIVATE:.*]] = acc.private varPtr(%[[BOX_ADDR]] : !fir.heap<!fir.array<?xi32>>) bounds(%{{.*}}) -> !fir.heap<!fir.array<?xi32>> {name = "a"}
! CHECK: acc.loop private({{.*}}@privatization_box_heap_Uxi32 -> %[[PRIVATE]] : !fir.heap<!fir.array<?xi32>>)
@@ -313,7 +315,7 @@ subroutine acc_private_pointer_array(a, n)
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>> {fir.bindc_name = "a"}, %arg1: !fir.ref<i32> {fir.bindc_name = "n"}) {
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %arg0 {fortran_attrs = #fir.var_attrs<pointer>, uniq_name = "_QFacc_private_pointer_arrayEa"} : (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>>) -> (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>>, !fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>>)
! CHECK: acc.parallel {
-! CHECK: %[[BOX:.*]] = fir.load %[[DECLA_A]]#1 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>>
+! CHECK: %[[BOX:.*]] = fir.load %[[DECLA_A]]#0 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xi32>>>>
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX]] : (!fir.box<!fir.ptr<!fir.array<?xi32>>>) -> !fir.ptr<!fir.array<?xi32>>
! CHECK: %[[PRIVATE:.*]] = acc.private varPtr(%[[BOX_ADDR]] : !fir.ptr<!fir.array<?xi32>>) bounds(%{{.*}}) -> !fir.ptr<!fir.array<?xi32>> {name = "a"}
! CHECK: acc.loop private({{.*}}@privatization_box_ptr_Uxi32 -> %[[PRIVATE]] : !fir.ptr<!fir.array<?xi32>>)
@@ -330,11 +332,12 @@ subroutine acc_private_dynamic_extent(a, n)
! CHECK-LABEL: func.func @_QPacc_private_dynamic_extent(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<?x?x2xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<i32> {fir.bindc_name = "n"}) {
-! CHECK: %[[DECL_N:.*]]:2 = hlfir.declare %arg1 {uniq_name = "_QFacc_private_dynamic_extentEn"} : (!fir.ref<i32>) -> (!fir.ref<i32>, !fir.ref<i32>)
-! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]](%16) {uniq_name = "_QFacc_private_dynamic_extentEa"} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
+! CHECK: %[[DECL_N:.*]]:2 = hlfir.declare %[[ARG1]] {uniq_name = "_QFacc_private_dynamic_extentEn"} : (!fir.ref<i32>) -> (!fir.ref<i32>, !fir.ref<i32>)
+! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {uniq_name = "_QFacc_private_dynamic_extentEa"} : (!fir.ref<!fir.array<?x?x2xi32>>, !fir.shape<3>) -> (!fir.box<!fir.array<?x?x2xi32>>, !fir.ref<!fir.array<?x?x2xi32>>)
! CHECK: acc.parallel {
-! CHECK: %[[PRIV:.*]] = acc.private varPtr(%[[DECL_A]]#1 : !fir.ref<!fir.array<?x?x2xi32>>) bounds(%{{.*}}, %{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<?x?x2xi32>> {name = "a"}
-! CHECK: acc.loop private({{.*}}@privatization_ref_UxUx2xi32 -> %[[PRIV]] : !fir.ref<!fir.array<?x?x2xi32>>)
+! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL_A]]#0 : (!fir.box<!fir.array<?x?x2xi32>>) -> !fir.ref<!fir.array<?x?x2xi32>>
+! CHECK: %[[PRIV:.*]] = acc.private varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?x?x2xi32>>) bounds(%{{.*}}, %{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<?x?x2xi32>> {name = "a"}
+! CHECK: acc.loop private({{.*}}@privatization_box_UxUx2xi32 -> %[[PRIV]] : !fir.ref<!fir.array<?x?x2xi32>>)
subroutine acc_firstprivate_assumed_shape(a, n)
integer :: a(:), i, n
@@ -364,7 +367,7 @@ subroutine acc_firstprivate_dynamic_extent(a, n)
end do
end subroutine
-! CHECK: acc.parallel firstprivate(@firstprivatization_ref_UxUx2xi32 -> %{{.*}} : !fir.ref<!fir.array<?x?x2xi32>>)
+! CHECK: acc.parallel firstprivate(@firstprivatization_box_UxUx2xi32 -> %{{.*}} : !fir.ref<!fir.array<?x?x2xi32>>)
module acc_declare_equivalent
integer, parameter :: n = 10
diff --git a/flang/test/Lower/OpenACC/acc-reduction.f90 b/flang/test/Lower/OpenACC/acc-reduction.f90
index 66a3153191723..6918bc1ec7d61 100644
--- a/flang/test/Lower/OpenACC/acc-reduction.f90
+++ b/flang/test/Lower/OpenACC/acc-reduction.f90
@@ -1,34 +1,35 @@
! This test checks lowering of OpenACC reduction clause.
-! RUN: bbc -fopenacc -emit-hlfir %s -o - | FileCheck %s --check-prefixes=CHECK
+! RUN: bbc -fopenacc -emit-hlfir %s -o - | FileCheck %s
-! CHECK-LABEL: acc.reduction.recipe @reduction_max_ref_UxUxf32 : !fir.ref<!fir.array<?x?xf32>> reduction_operator <max> init {
-! CHECK: ^bb0(%[[ARG0:.*]]: !fir.ref<!fir.array<?x?xf32>>, %[[ARG1:.*]]: index, %[[ARG2:.*]]: index):
+! CHECK-LABEL: acc.reduction.recipe @reduction_max_box_UxUxf32 : !fir.box<!fir.array<?x?xf32>> reduction_operator <max> init {
+! CHECK: ^bb0(%[[ARG0:.*]]: !fir.box<!fir.array<?x?xf32>>):
! CHECK: %[[CST:.*]] = arith.constant -1.401300e-45 : f32
-! CHECK: %[[SHAPE:.*]] = fir.shape %arg1, %arg2 : (index, index) -> !fir.shape<2>
-! CHECK: %[[TEMP:.*]] = fir.alloca !fir.array<?x?xf32>, %arg1, %arg2
-! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[TEMP]](%[[SHAPE]]) {uniq_name = "acc.reduction.init"} : (!fir.ref<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.ref<!fir.array<?x?xf32>>)
+! CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %[[ARG0]], %c0{{.*}} : (!fir.box<!fir.array<?x?xf32>>, index) -> (index, index, index)
+! CHECK: %[[DIMS1:.*]]:3 = fir.box_dims %[[ARG0]], %c1 : (!fir.box<!fir.array<?x?xf32>>, index) -> (index, index, index)
+! CHECK: %[[SHAPE:.*]] = fir.shape %[[DIMS0]]#1, %[[DIMS1]]#1 : (index, index) -> !fir.shape<2>
+! CHECK: %[[TEMP:.*]] = fir.allocmem !fir.array<?x?xf32>, %[[DIMS0]]#1, %[[DIMS1]]#1 {bindc_name = ".tmp", uniq_name = ""}
+! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[TEMP]](%[[SHAPE]]) {uniq_name = ".tmp"} : (!fir.heap<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.heap<!fir.array<?x?xf32>>)
! CHECK: hlfir.assign %[[CST]] to %[[DECL]]#0 : f32, !fir.box<!fir.array<?x?xf32>>
! CHECK: acc.yield %[[DECL]]#0 : !fir.box<!fir.array<?x?xf32>>
! CHECK: } combiner {
-! CHECK: ^bb0(%[[V1:.*]]: !fir.ref<!fir.array<?x?xf32>>, %[[V2:.*]]: !fir.ref<!fir.array<?x?xf32>>, %[[LB0:.*]]: index, %[[UB0:.*]]: index, %[[STEP0:.*]]: index, %[[LB1:.*]]: index, %[[UB1:.*]]: index, %[[STEP1:.*]]: index):
+! CHECK: ^bb0(%[[V1:.*]]: !fir.box<!fir.array<?x?xf32>>, %[[V2:.*]]: !fir.box<!fir.array<?x?xf32>>, %[[LB0:.*]]: index, %[[UB0:.*]]: index, %[[STEP0:.*]]: index, %[[LB1:.*]]: index, %[[UB1:.*]]: index, %[[STEP1:.*]]: index):
+
! CHECK: %[[SHAPE:.*]] = fir.shape %{{.*}}, %{{.*}} : (index, index) -> !fir.shape<2>
-! CHECK: %[[DECL_V1:.*]]:2 = hlfir.declare %[[V1]](%[[SHAPE]]) {uniq_name = ""} : (!fir.ref<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.ref<!fir.array<?x?xf32>>)
-! CHECK: %[[DECL_V2:.*]]:2 = hlfir.declare %[[V2]](%[[SHAPE]]) {uniq_name = ""} : (!fir.ref<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.ref<!fir.array<?x?xf32>>)
-! CHECK: %[[DES_V1:.*]] = hlfir.designate %[[DECL_V1]]#0 (%arg2:%arg3:%arg4, %arg5:%arg6:%arg7) shape %10 : (!fir.box<!fir.array<?x?xf32>>, index, index, index, index, index, index, !fir.shape<2>) -> !fir.box<!fir.array<?x?xf32>>
-! CHECK: %[[DES_V2:.*]] = hlfir.designate %[[DECL_V2]]#0 (%arg2:%arg3:%arg4, %arg5:%arg6:%arg7) shape %10 : (!fir.box<!fir.array<?x?xf32>>, index, index, index, index, index, index, !fir.shape<2>) -> !fir.box<!fir.array<?x?xf32>>
+! CHECK: %[[DES_V1:.*]] = hlfir.designate %[[V1]] (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?xf32>>, index, index, index, index, index, index, !fir.shape<2>) -> !fir.box<!fir.array<?x?xf32>>
+! CHECK: %[[DES_V2:.*]] = hlfir.designate %[[V2]] (%[[LB0]]:%[[UB0]]:%[[STEP0]], %[[LB1]]:%[[UB1]]:%[[STEP1]]) shape %[[SHAPE]] : (!fir.box<!fir.array<?x?xf32>>, index, index, index, index, index, index, !fir.shape<2>) -> !fir.box<!fir.array<?x?xf32>>
! CHECK: %[[ELEMENTAL:.*]] = hlfir.elemental %[[SHAPE]] unordered : (!fir.shape<2>) -> !hlfir.expr<?x?xf32> {
! CHECK: ^bb0(%[[ARG0:.*]]: index, %[[ARG1:.*]]: index):
-! CHECK: %[[D1:.*]] = hlfir.designate %13 (%[[ARG0]], %[[ARG1]]) : (!fir.box<!fir.array<?x?xf32>>, index, index) -> !fir.ref<f32>
-! CHECK: %[[D2:.*]] = hlfir.designate %14 (%[[ARG0]], %[[ARG1]]) : (!fir.box<!fir.array<?x?xf32>>, index, index) -> !fir.ref<f32>
+! CHECK: %[[D1:.*]] = hlfir.designate %[[DES_V1]] (%[[ARG0]], %[[ARG1]]) : (!fir.box<!fir.array<?x?xf32>>, index, index) -> !fir.ref<f32>
+! CHECK: %[[D2:.*]] = hlfir.designate %[[DES_V2]] (%[[ARG0]], %[[ARG1]]) : (!fir.box<!fir.array<?x?xf32>>, index, index) -> !fir.ref<f32>
! CHECK: %[[LOAD1:.*]] = fir.load %[[D1]] : !fir.ref<f32>
! CHECK: %[[LOAD2:.*]] = fir.load %[[D2]] : !fir.ref<f32>
! CHECK: %[[CMP:.*]] = arith.cmpf ogt, %[[LOAD1]], %[[LOAD2]] {{.*}} : f32
! CHECK: %[[SELECT:.*]] = arith.select %[[CMP]], %[[LOAD1]], %[[LOAD2]] : f32
! CHECK: hlfir.yield_element %[[SELECT]] : f32
! CHECK: }
-! CHECK: hlfir.assign %[[ELEMENTAL]] to %[[DECL_V1]]#0 : !hlfir.expr<?x?xf32>, !fir.box<!fir.array<?x?xf32>>
-! CHECK: acc.yield %[[V1]] : !fir.ref<!fir.array<?x?xf32>>
+! CHECK: hlfir.assign %[[ELEMENTAL]] to %[[V1]] : !hlfir.expr<?x?xf32>, !fir.box<!fir.array<?x?xf32>>
+! CHECK: acc.yield %[[V1]] : !fir.box<!fir.array<?x?xf32>>
! CHECK: }
! CHECK-LABEL: acc.reduction.recipe @reduction_max_box_ptr_Uxf32 : !fir.box<!fir.ptr<!fir.array<?xf32>>> reduction_operator <max> init {
@@ -703,7 +704,7 @@ subroutine acc_reduction_add_int(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_int(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<i32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_ref_i32 -> %[[RED_B]] : !fir.ref<i32>)
subroutine acc_reduction_add_int_array_1d(a, b)
@@ -719,7 +720,7 @@ subroutine acc_reduction_add_int_array_1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_int_array_1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_section_ext100_ref_100xi32 -> %[[RED_B]] : !fir.ref<!fir.array<100xi32>>)
subroutine acc_reduction_add_int_array_2d(a, b)
@@ -737,7 +738,7 @@ subroutine acc_reduction_add_int_array_2d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_int_array_2d(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<100x10xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100x10xi32>> {fir.bindc_name = "b"}) {
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100x10xi32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xi32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100x10xi32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_section_ext100xext10_ref_100x10xi32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100x10xi32>>)
! CHECK: } attributes {collapse = [2]{{.*}}
@@ -758,7 +759,7 @@ subroutine acc_reduction_add_int_array_3d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_int_array_3d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100x10x2xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100x10x2xi32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100x10x2xi32>>) bounds(%{{.*}}, %{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10x2xi32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100x10x2xi32>>) bounds(%{{.*}}, %{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10x2xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_section_ext100xext10xext2_ref_100x10x2xi32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100x10x2xi32>>)
! CHECK: } attributes {collapse = [3]{{.*}}
@@ -775,7 +776,7 @@ subroutine acc_reduction_add_float(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_float(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<f32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_ref_f32 -> %[[RED_B]] : !fir.ref<f32>)
subroutine acc_reduction_add_float_array_1d(a, b)
@@ -791,7 +792,7 @@ subroutine acc_reduction_add_float_array_1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_add_float_array_1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_add_section_ext100_ref_100xf32 -> %[[RED_B]] : !fir.ref<!fir.array<100xf32>>)
subroutine acc_reduction_mul_int(a, b)
@@ -807,7 +808,7 @@ subroutine acc_reduction_mul_int(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_mul_int(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<i32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_mul_ref_i32 -> %[[RED_B]] : !fir.ref<i32>)
subroutine acc_reduction_mul_int_array_1d(a, b)
@@ -823,7 +824,7 @@ subroutine acc_reduction_mul_int_array_1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_mul_int_array_1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_mul_section_ext100_ref_100xi32 -> %[[RED_B]] : !fir.ref<!fir.array<100xi32>>)
subroutine acc_reduction_mul_float(a, b)
@@ -839,7 +840,7 @@ subroutine acc_reduction_mul_float(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_mul_float(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<f32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_mul_ref_f32 -> %[[RED_B]] : !fir.ref<f32>)
subroutine acc_reduction_mul_float_array_1d(a, b)
@@ -855,7 +856,7 @@ subroutine acc_reduction_mul_float_array_1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_mul_float_array_1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_mul_section_ext100_ref_100xf32 -> %[[RED_B]] : !fir.ref<!fir.array<100xf32>>)
subroutine acc_reduction_min_int(a, b)
@@ -871,7 +872,7 @@ subroutine acc_reduction_min_int(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_min_int(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<i32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_min_ref_i32 -> %[[RED_B]] : !fir.ref<i32>)
subroutine acc_reduction_min_int_array_1d(a, b)
@@ -887,7 +888,7 @@ subroutine acc_reduction_min_int_array_1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_min_int_array_1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_min_section_ext100_ref_100xi32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100xi32>>)
subroutine acc_reduction_min_float(a, b)
@@ -903,7 +904,7 @@ subroutine acc_reduction_min_float(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_min_float(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<f32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_min_ref_f32 -> %[[RED_B]] : !fir.ref<f32>)
subroutine acc_reduction_min_float_array2d(a, b)
@@ -921,7 +922,7 @@ subroutine acc_reduction_min_float_array2d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_min_float_array2d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100x10xf32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100x10xf32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xf32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xf32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_min_section_ext100xext10_ref_100x10xf32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100x10xf32>>)
! CHECK: attributes {collapse = [2]{{.*}}
@@ -938,7 +939,7 @@ subroutine acc_reduction_max_int(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_max_int(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xi32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<i32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<i32>) -> !fir.ref<i32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_max_ref_i32 -> %[[RED_B]] : !fir.ref<i32>)
subroutine acc_reduction_max_int_array2d(a, b)
@@ -956,7 +957,7 @@ subroutine acc_reduction_max_int_array2d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_max_int_array2d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100x10xi32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100x10xi32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100x10xi32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xi32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100x10xi32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<100x10xi32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_max_section_ext100xext10_ref_100x10xi32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100x10xi32>>)
subroutine acc_reduction_max_float(a, b)
@@ -972,7 +973,7 @@ subroutine acc_reduction_max_float(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_max_float(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[B:.*]]: !fir.ref<f32> {fir.bindc_name = "b"})
! CHECK: %[[DECLB:.*]]:2 = hlfir.declare %[[B]]
-! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#1 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
+! CHECK: %[[RED_B:.*]] = acc.reduction varPtr(%[[DECLB]]#0 : !fir.ref<f32>) -> !fir.ref<f32> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_max_ref_f32 -> %[[RED_B]] : !fir.ref<f32>)
subroutine acc_reduction_max_float_array1d(a, b)
@@ -988,7 +989,7 @@ subroutine acc_reduction_max_float_array1d(a, b)
! CHECK-LABEL: func.func @_QPacc_reduction_max_float_array1d(
! CHECK-SAME: %{{.*}}: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "a"}, %[[ARG1:.*]]: !fir.ref<!fir.array<100xf32>> {fir.bindc_name = "b"})
! CHECK: %[[DECLARG1:.*]]:2 = hlfir.declare %[[ARG1]]
-! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#1 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
+! CHECK: %[[RED_ARG1:.*]] = acc.reduction varPtr(%[[DECLARG1]]#0 : !fir.ref<!fir.array<100xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<100xf32>> {name = "b"}
! CHECK: acc.loop {{.*}} reduction(@reduction_max_section_ext100_ref_100xf32 -> %[[RED_ARG1]] : !fir.ref<!fir.array<100xf32>>)
subroutine acc_reduction_iand()
@@ -1030,7 +1031,7 @@ subroutine acc_reduction_and()
! CHECK-LABEL: func.func @_QPacc_reduction_and()
! CHECK: %[[L:.*]] = fir.alloca !fir.logical<4> {bindc_name = "l", uniq_name = "_QFacc_reduction_andEl"}
! CHECK: %[[DECLL:.*]]:2 = hlfir.declare %[[L]]
-! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[DECLL]]#1 : !fir.ref<!fir.logical<4>>) -> !fir.ref<!fir.logical<4>> {name = "l"}
+! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[DECLL]]#0 : !fir.ref<!fir.logical<4>>) -> !fir.ref<!fir.logical<4>> {name = "l"}
! CHECK: acc.parallel reduction(@reduction_land_ref_l32 -> %[[RED]] : !fir.ref<!fir.logical<4>>)
subroutine acc_reduction_or()
@@ -1093,7 +1094,7 @@ subroutine acc_reduction_add_alloc()
! CHECK-LABEL: func.func @_QPacc_reduction_add_alloc()
! CHECK: %[[ALLOCA:.*]] = fir.alloca !fir.box<!fir.heap<i32>> {bindc_name = "i", uniq_name = "_QFacc_reduction_add_allocEi"}
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ALLOCA]]
-! CHECK: %[[LOAD:.*]] = fir.load %[[DECL]]#1 : !fir.ref<!fir.box<!fir.heap<i32>>>
+! CHECK: %[[LOAD:.*]] = fir.load %[[DECL]]#0 : !fir.ref<!fir.box<!fir.heap<i32>>>
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[LOAD]] : (!fir.box<!fir.heap<i32>>) -> !fir.heap<i32>
! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[BOX_ADDR]] : !fir.heap<i32>) -> !fir.heap<i32> {name = "i"}
! CHECK: acc.parallel reduction(@reduction_add_heap_i32 -> %[[RED]] : !fir.heap<i32>)
@@ -1107,7 +1108,7 @@ subroutine acc_reduction_add_pointer(i)
! CHECK-LABEL: func.func @_QPacc_reduction_add_pointer(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.ptr<i32>>> {fir.bindc_name = "i"})
! CHECK: %[[DECLARG0:.*]]:2 = hlfir.declare %[[ARG0]]
-! CHECK: %[[LOAD:.*]] = fir.load %[[DECLARG0]]#1 : !fir.ref<!fir.box<!fir.ptr<i32>>>
+! CHECK: %[[LOAD:.*]] = fir.load %[[DECLARG0]]#0 : !fir.ref<!fir.box<!fir.ptr<i32>>>
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[LOAD]] : (!fir.box<!fir.ptr<i32>>) -> !fir.ptr<i32>
! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[BOX_ADDR]] : !fir.ptr<i32>) -> !fir.ptr<i32> {name = "i"}
! CHECK: acc.parallel reduction(@reduction_add_ptr_i32 -> %[[RED]] : !fir.ptr<i32>)
@@ -1126,7 +1127,7 @@ subroutine acc_reduction_add_static_slice(a)
! CHECK: %[[LB:.*]] = arith.constant 10 : index
! CHECK: %[[UB:.*]] = arith.constant 19 : index
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%[[LB]] : index) upperbound(%[[UB]] : index) extent(%[[C100]] : index) stride(%[[C1]] : index) startIdx(%[[C1]] : index)
-! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[DECLARG0]]#1 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a(11:20)"}
+! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[DECLARG0]]#0 : !fir.ref<!fir.array<100xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<100xi32>> {name = "a(11:20)"}
! CHECK: acc.parallel reduction(@reduction_add_section_lb10.ub19_ref_100xi32 -> %[[RED]] : !fir.ref<!fir.array<100xi32>>)
subroutine acc_reduction_add_dynamic_extent_add(a)
@@ -1163,7 +1164,7 @@ subroutine acc_reduction_add_dynamic_extent_add_with_section(a)
! CHECK-SAME: %[[ARG0:.*]]: !fir.box<!fir.array<?xi32>> {fir.bindc_name = "a"})
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]] {uniq_name = "_QFacc_reduction_add_dynamic_extent_add_with_sectionEa"} : (!fir.box<!fir.array<?xi32>>) -> (!fir.box<!fir.array<?xi32>>, !fir.box<!fir.array<?xi32>>)
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c1{{.*}} : index) upperbound(%c3{{.*}} : index) extent(%{{.*}}#1 : index) stride(%{{.*}}#2 : index) startIdx(%{{.*}} : index) {strideInBytes = true}
-! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL]]#1 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
+! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[DECL]]#0 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[BOX_ADDR]] : !fir.ref<!fir.array<?xi32>>) bounds(%[[BOUND]]) -> !fir.ref<!fir.array<?xi32>> {name = "a(2:4)"}
! CHECK: acc.parallel reduction(@reduction_add_section_lb1.ub3_box_Uxi32 -> %[[RED]] : !fir.ref<!fir.array<?xi32>>)
@@ -1176,7 +1177,7 @@ subroutine acc_reduction_add_allocatable(a)
! CHECK-LABEL: func.func @_QPacc_reduction_add_allocatable(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>> {fir.bindc_name = "a"})
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]] {fortran_attrs = #fir.var_attrs<allocatable>, uniq_name = "_QFacc_reduction_add_allocatableEa"} : (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>)
-! CHECK: %[[BOX:.*]] = fir.load %[[DECL]]#1 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
+! CHECK: %[[BOX:.*]] = fir.load %[[DECL]]#0 : !fir.ref<!fir.box<!fir.heap<!fir.array<?xf32>>>>
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c0{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}}#1 : index) stride(%{{.*}}#2 : index) startIdx(%{{.*}}#0 : index) {strideInBytes = true}
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX]] : (!fir.box<!fir.heap<!fir.array<?xf32>>>) -> !fir.heap<!fir.array<?xf32>>
! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[BOX_ADDR]] : !fir.heap<!fir.array<?xf32>>) bounds(%6) -> !fir.heap<!fir.array<?xf32>> {name = "a"}
@@ -1191,7 +1192,7 @@ subroutine acc_reduction_add_pointer_array(a)
! CHECK-LABEL: func.func @_QPacc_reduction_add_pointer_array(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>> {fir.bindc_name = "a"})
! CHECK: %[[DECL:.*]]:2 = hlfir.declare %[[ARG0]] {fortran_attrs = #fir.var_attrs<pointer>, uniq_name = "_QFacc_reduction_add_pointer_arrayEa"} : (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>) -> (!fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>, !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>)
-! CHECK: %[[BOX:.*]] = fir.load %[[DECL]]#1 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
+! CHECK: %[[BOX:.*]] = fir.load %[[DECL]]#0 : !fir.ref<!fir.box<!fir.ptr<!fir.array<?xf32>>>>
! CHECK: %[[BOUND:.*]] = acc.bounds lowerbound(%c0{{.*}} : index) upperbound(%{{.*}} : index) extent(%{{.*}}#1 : index) stride(%{{.*}}#2 : index) startIdx(%{{.*}}#0 : index) {strideInBytes = true}
! CHECK: %[[BOX_ADDR:.*]] = fir.box_addr %[[BOX]] : (!fir.box<!fir.ptr<!fir.array<?xf32>>>) -> !fir.ptr<!fir.array<?xf32>>
! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[BOX_ADDR]] : !fir.ptr<!fir.array<?xf32>>) bounds(%[[BOUND]]) -> !fir.ptr<!fir.array<?xf32>> {name = "a"}
@@ -1207,5 +1208,6 @@ subroutine acc_reduction_max_dynamic_extent_max(a, n)
! CHECK-LABEL: func.func @_QPacc_reduction_max_dynamic_extent_max(
! CHECK-SAME: %[[ARG0:.*]]: !fir.ref<!fir.array<?x?xf32>> {fir.bindc_name = "a"}, %{{.*}}: !fir.ref<i32> {fir.bindc_name = "n"})
! CHECK: %[[DECL_A:.*]]:2 = hlfir.declare %[[ARG0]](%{{.*}}) {uniq_name = "_QFacc_reduction_max_dynamic_extent_maxEa"} : (!fir.ref<!fir.array<?x?xf32>>, !fir.shape<2>) -> (!fir.box<!fir.array<?x?xf32>>, !fir.ref<!fir.array<?x?xf32>>)
-! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[DECL_A]]#1 : !fir.ref<!fir.array<?x?xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<?x?xf32>> {name = "a"}
-! CHECK: acc.parallel reduction(@reduction_max_ref_UxUxf32 -> %[[RED]] : !fir.ref<!fir.array<?x?xf32>>)
+! CHECK: %[[ADDR:.*]] = fir.box_addr %[[DECL_A]]#0 : (!fir.box<!fir.array<?x?xf32>>) -> !fir.ref<!fir.array<?x?xf32>>
+! CHECK: %[[RED:.*]] = acc.reduction varPtr(%[[ADDR]] : !fir.ref<!fir.array<?x?xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<?x?xf32>> {name = "a"}
+! CHECK: acc.parallel reduction(@reduction_max_box_UxUxf32 -> %[[RED]] : !fir.ref<!fir.array<?x?xf32>>)
diff --git a/flang/test/Lower/OpenACC/acc-serial-loop.f90 b/flang/test/Lower/OpenACC/acc-serial-loop.f90
index 9c0dbff0d7dac..2869077bcb1ff 100644
--- a/flang/test/Lower/OpenACC/acc-serial-loop.f90
+++ b/flang/test/Lower/OpenACC/acc-serial-loop.f90
@@ -232,40 +232,40 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc serial loop copy(a) copy(b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
!$acc serial loop copyin(a) copyin(readonly: b)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -278,24 +278,24 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
! CHECK: acc.serial dataOperands(%[[CREATE_A]], %[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) {name = "b"}
!$acc serial loop create(b) create(zero: a)
DO i = 1, n
a(i) = b(i)
END DO
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "a"}
! CHECK: acc.serial dataOperands(%[[CREATE_B]], %[[CREATE_A]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -310,8 +310,8 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[NOCREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[NOCREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.serial dataOperands(%[[NOCREATE_A]], %[[NOCREATE_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -324,8 +324,8 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.serial dataOperands(%[[PRESENT_A]], %[[PRESENT_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -338,8 +338,8 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_B:.*]] = acc.deviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.serial dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_B]] : !fir.ref<!fir.array<10xf32>>, !fir.ref<!fir.array<10xf32>>) {
! CHECK: acc.loop {{.*}} {
! CHECK: acc.yield
@@ -352,10 +352,10 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_F:.*]] = fir.load %[[DECLF]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_F:.*]] = fir.box_addr %[[BOX_F]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_F:.*]] = acc.attach varPtr(%[[BOX_ADDR_F]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "f"}
-! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_G:.*]] = fir.load %[[DECLG]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_G:.*]] = fir.box_addr %[[BOX_G]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_G:.*]] = acc.attach varPtr(%[[BOX_ADDR_G]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "g"}
! CHECK: acc.serial dataOperands(%[[ATTACH_F]], %[[ATTACH_G]] : !fir.ptr<f32>, !fir.ptr<f32>) {
@@ -370,9 +370,9 @@ subroutine acc_serial_loop
a(i) = b(i)
END DO
-! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
+! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "b"}
! CHECK: acc.serial firstprivate(@firstprivatization_section_ext10_ref_10xf32 -> %[[ACC_FPRIVATE_B]] : !fir.ref<!fir.array<10xf32>>) {
-! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
+! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10xf32>>) bounds(%{{.*}}) -> !fir.ref<!fir.array<10xf32>> {name = "a"}
! CHECK: acc.loop private({{.*}}@privatization_ref_10xf32 -> %[[ACC_PRIVATE_A]] : !fir.ref<!fir.array<10xf32>>)
! CHECK-NOT: fir.do_loop
! CHECK: acc.yield
diff --git a/flang/test/Lower/OpenACC/acc-serial.f90 b/flang/test/Lower/OpenACC/acc-serial.f90
index d0fa9436be14a..c6fe6c3d58fe5 100644
--- a/flang/test/Lower/OpenACC/acc-serial.f90
+++ b/flang/test/Lower/OpenACC/acc-serial.f90
@@ -156,35 +156,35 @@ subroutine acc_serial
!$acc serial copy(a, b, c)
!$acc end serial
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc serial copy(a) copy(b) copy(c)
!$acc end serial
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copy>, name = "c"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
-! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "a"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "b"}
+! CHECK: acc.copyout accPtr(%[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_copy>, name = "c"}
!$acc serial copyin(a) copyin(readonly: b, c)
!$acc end serial
-! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
-! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
+! CHECK: %[[COPYIN_A:.*]] = acc.copyin varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[COPYIN_B:.*]] = acc.copyin varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "b"}
+! CHECK: %[[COPYIN_C:.*]] = acc.copyin varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyin_readonly>, name = "c"}
! CHECK: acc.serial dataOperands(%[[COPYIN_A]], %[[COPYIN_B]], %[[COPYIN_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -192,22 +192,22 @@ subroutine acc_serial
!$acc serial copyout(a) copyout(zero: b) copyout(c)
!$acc end serial
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_copyout>, name = "c"}
! CHECK: acc.serial dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
-! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
-! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
-! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
+! CHECK: acc.copyout accPtr(%[[CREATE_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a"}
+! CHECK: acc.copyout accPtr(%[[CREATE_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "b"}
+! CHECK: acc.copyout accPtr(%[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "c"}
!$acc serial create(a, b) create(zero: c)
!$acc end serial
-! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[CREATE_A:.*]] = acc.create varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[CREATE_B:.*]] = acc.create varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.serial dataOperands(%[[CREATE_A]], %[[CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -218,9 +218,9 @@ subroutine acc_serial
!$acc serial no_create(a, b) create(zero: c)
!$acc end serial
-! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
+! CHECK: %[[NO_CREATE_A:.*]] = acc.nocreate varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[NO_CREATE_B:.*]] = acc.nocreate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[CREATE_C:.*]] = acc.create varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_create_zero>, name = "c"}
! CHECK: acc.serial dataOperands(%[[NO_CREATE_A]], %[[NO_CREATE_B]], %[[CREATE_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -228,9 +228,9 @@ subroutine acc_serial
!$acc serial present(a, b, c)
!$acc end serial
-! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[PRESENT_A:.*]] = acc.present varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[PRESENT_B:.*]] = acc.present varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[PRESENT_C:.*]] = acc.present varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.serial dataOperands(%[[PRESENT_A]], %[[PRESENT_B]], %[[PRESENT_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -238,8 +238,8 @@ subroutine acc_serial
!$acc serial deviceptr(a) deviceptr(c)
!$acc end serial
-! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[DEVICEPTR_A:.*]] = acc.deviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[DEVICEPTR_C:.*]] = acc.deviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.serial dataOperands(%[[DEVICEPTR_A]], %[[DEVICEPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
@@ -247,10 +247,10 @@ subroutine acc_serial
!$acc serial attach(d, e)
!$acc end serial
-! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_D:.*]] = fir.load %[[DECLD]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_D:.*]] = fir.box_addr %[[BOX_D]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_D:.*]] = acc.attach varPtr(%[[BOX_ADDR_D]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "d"}
-! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#1 : !fir.ref<!fir.box<!fir.ptr<f32>>>
+! CHECK: %[[BOX_E:.*]] = fir.load %[[DECLE]]#0 : !fir.ref<!fir.box<!fir.ptr<f32>>>
! CHECK: %[[BOX_ADDR_E:.*]] = fir.box_addr %[[BOX_E]] : (!fir.box<!fir.ptr<f32>>) -> !fir.ptr<f32>
! CHECK: %[[ATTACH_E:.*]] = acc.attach varPtr(%[[BOX_ADDR_E]] : !fir.ptr<f32>) -> !fir.ptr<f32> {name = "e"}
! CHECK: acc.serial dataOperands(%[[ATTACH_D]], %[[ATTACH_E]] : !fir.ptr<f32>, !fir.ptr<f32>) {
@@ -262,9 +262,9 @@ subroutine acc_serial
!$acc serial private(a) firstprivate(b) private(c)
!$acc end serial
-! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
-! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
-! CHECK: %[[ACC_PRIVATE_C:.*]] = acc.private varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
+! CHECK: %[[ACC_PRIVATE_A:.*]] = acc.private varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "a"}
+! CHECK: %[[ACC_FPRIVATE_B:.*]] = acc.firstprivate varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "b"}
+! CHECK: %[[ACC_PRIVATE_C:.*]] = acc.private varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c"}
! CHECK: acc.serial firstprivate(@firstprivatization_section_ext10xext10_ref_10x10xf32 -> %[[ACC_FPRIVATE_B]] : !fir.ref<!fir.array<10x10xf32>>) private(@privatization_ref_10x10xf32 -> %[[ACC_PRIVATE_A]] : !fir.ref<!fir.array<10x10xf32>>, @privatization_ref_10x10xf32 -> %[[ACC_PRIVATE_C]] : !fir.ref<!fir.array<10x10xf32>>) {
! CHECK: acc.yield
! CHECK-NEXT: }{{$}}
diff --git a/flang/test/Lower/OpenACC/acc-update.f90 b/flang/test/Lower/OpenACC/acc-update.f90
index f42ae1356664b..0c30ea5d44c61 100644
--- a/flang/test/Lower/OpenACC/acc-update.f90
+++ b/flang/test/Lower/OpenACC/acc-update.f90
@@ -15,98 +15,98 @@ subroutine acc_update
! CHECK: %[[DECLC:.*]]:2 = hlfir.declare %[[C]]
!$acc update host(a)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) if_present
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) attributes {ifPresent}{{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update self(a)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_self>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_self>, name = "a", structured = false}
! CHECK: acc.update dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_update_self>, name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {dataClause = #acc<data_clause acc_update_self>, name = "a", structured = false}
!$acc update host(a) if(.true.)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: %[[IF1:.*]] = arith.constant true
! CHECK: acc.update if(%[[IF1]]) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) if(ifCondition)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: %[[IFCOND:.*]] = fir.load %{{.*}} : !fir.ref<!fir.logical<4>>
! CHECK: %[[IF2:.*]] = fir.convert %[[IFCOND]] : (!fir.logical<4>) -> i1
! CHECK: acc.update if(%[[IF2]]) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) host(b) host(c)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
-! CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "b", structured = false}
-! CHECK: %[[DEVPTR_C:.*]] = acc.getdeviceptr varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "c", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "b", structured = false}
+! CHECK: %[[DEVPTR_C:.*]] = acc.getdeviceptr varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "c", structured = false}
! CHECK: acc.update dataOperands(%[[DEVPTR_A]], %[[DEVPTR_B]], %[[DEVPTR_C]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "b", structured = false}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "c", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "b", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_C]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "c", structured = false}
!$acc update host(a) host(b) device(c)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
-! CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "b", structured = false}
-! CHECK: %[[DEVPTR_C:.*]] = acc.update_device varPtr(%[[DECLC]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_B:.*]] = acc.getdeviceptr varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "b", structured = false}
+! CHECK: %[[DEVPTR_C:.*]] = acc.update_device varPtr(%[[DECLC]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {name = "c", structured = false}
! CHECK: acc.update dataOperands(%[[DEVPTR_C]], %[[DEVPTR_A]], %[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>, !fir.ref<!fir.array<10x10xf32>>){{$}}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
-! CHECK: acc.update_host accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "b", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_B]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLB]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "b", structured = false}
!$acc update host(a) async
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update async() dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) wait
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update wait dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) async wait
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update async() wait dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) async(1)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: [[ASYNC1:%.*]] = arith.constant 1 : i32
! CHECK: acc.update async([[ASYNC1]] : i32) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) async(async)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: [[ASYNC2:%.*]] = fir.load %{{.*}} : !fir.ref<i32>
! CHECK: acc.update async([[ASYNC2]] : i32) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) wait(1)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: [[WAIT1:%.*]] = arith.constant 1 : i32
! CHECK: acc.update wait({[[WAIT1]] : i32}) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) wait(queues: 1, 2)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: [[WAIT2:%.*]] = arith.constant 1 : i32
! CHECK: [[WAIT3:%.*]] = arith.constant 2 : i32
! CHECK: acc.update wait({[[WAIT2]] : i32, [[WAIT3]] : i32}) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) wait(devnum: 1: queues: 1, 2)
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update wait({devnum: %c1{{.*}} : i32, %c1{{.*}} : i32, %c2{{.*}} : i32}) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
!$acc update host(a) device_type(host, nvidia) async
-! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
+! CHECK: %[[DEVPTR_A:.*]] = acc.getdeviceptr varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) -> !fir.ref<!fir.array<10x10xf32>> {dataClause = #acc<data_clause acc_update_host>, name = "a", structured = false}
! CHECK: acc.update async([#acc.device_type<host>, #acc.device_type<nvidia>]) dataOperands(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>)
-! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#1 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
+! CHECK: acc.update_host accPtr(%[[DEVPTR_A]] : !fir.ref<!fir.array<10x10xf32>>) bounds(%{{.*}}, %{{.*}}) to varPtr(%[[DECLA]]#0 : !fir.ref<!fir.array<10x10xf32>>) {name = "a", structured = false}
end subroutine acc_update
diff --git a/flang/test/Lower/OpenMP/array-bounds.f90 b/flang/test/Lower/OpenMP/array-bounds.f90
index 0fb5af188fa93..92c0c5307f0a1 100644
--- a/flang/test/Lower/OpenMP/array-bounds.f90
+++ b/flang/test/Lower/OpenMP/array-bounds.f90
@@ -15,12 +15,12 @@
!HOST: %[[C2:.*]] = arith.constant 1 : index
!HOST: %[[C3:.*]] = arith.constant 4 : index
!HOST: %[[BOUNDS0:.*]] = omp.bounds lower_bound(%[[C2]] : index) upper_bound(%[[C3]] : index) extent(%[[C10]] : index) stride(%[[C1]] : index) start_idx(%[[C1]] : index)
-!HOST: %[[MAP0:.*]] = omp.map_info var_ptr(%[[READ_DECL]]#1 : !fir.ref<!fir.array<10xi32>>, !fir.array<10xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS0]]) -> !fir.ref<!fir.array<10xi32>> {name = "sp_read(2:5)"}
+!HOST: %[[MAP0:.*]] = omp.map_info var_ptr(%[[READ_DECL]]#0 : !fir.ref<!fir.array<10xi32>>, !fir.array<10xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS0]]) -> !fir.ref<!fir.array<10xi32>> {name = "sp_read(2:5)"}
!HOST: %[[C4:.*]] = arith.constant 1 : index
!HOST: %[[C5:.*]] = arith.constant 1 : index
!HOST: %[[C6:.*]] = arith.constant 4 : index
!HOST: %[[BOUNDS1:.*]] = omp.bounds lower_bound(%[[C5]] : index) upper_bound(%[[C6]] : index) extent(%[[C10_0]] : index) stride(%[[C4]] : index) start_idx(%[[C4]] : index)
-!HOST: %[[MAP1:.*]] = omp.map_info var_ptr(%[[WRITE_DECL]]#1 : !fir.ref<!fir.array<10xi32>>, !fir.array<10xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS1]]) -> !fir.ref<!fir.array<10xi32>> {name = "sp_write(2:5)"}
+!HOST: %[[MAP1:.*]] = omp.map_info var_ptr(%[[WRITE_DECL]]#0 : !fir.ref<!fir.array<10xi32>>, !fir.array<10xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS1]]) -> !fir.ref<!fir.array<10xi32>> {name = "sp_write(2:5)"}
!HOST: omp.target map_entries(%[[MAP0]] -> %{{.*}}, %[[MAP1]] -> %{{.*}}, {{.*}} -> {{.*}} : !fir.ref<!fir.array<10xi32>>, !fir.ref<!fir.array<10xi32>>, !fir.ref<i32>) {
subroutine read_write_section()
@@ -43,13 +43,13 @@ module assumed_array_routines
!HOST: %[[ARG0_DECL:.*]]:2 = hlfir.declare %[[ARG0]] {fortran_attrs = #fir.var_attrs<intent_inout>, uniq_name = "_QMassumed_array_routinesFassumed_shape_arrayEarr_read_write"} : (!fir.box<!fir.array<?xi32>>) -> (!fir.box<!fir.array<?xi32>>, !fir.box<!fir.array<?xi32>>)
!HOST: %[[C0:.*]] = arith.constant 1 : index
!HOST: %[[C1:.*]] = arith.constant 0 : index
-!HOST: %[[DIMS0:.*]]:3 = fir.box_dims %[[ARG0_DECL]]#1, %[[C1]] : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index)
+!HOST: %[[DIMS0:.*]]:3 = fir.box_dims %[[ARG0_DECL]]#0, %[[C1]] : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index)
!HOST: %[[C3:.*]] = arith.constant 1 : index
!HOST: %[[C4:.*]] = arith.constant 4 : index
!HOST: %[[C0_1:.*]] = arith.constant 0 : index
!HOST: %[[DIMS1:.*]]:3 = fir.box_dims %[[ARG0_DECL]]#1, %[[C0_1]] : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index)
!HOST: %[[BOUNDS:.*]] = omp.bounds lower_bound(%[[C3]] : index) upper_bound(%[[C4]] : index) extent(%[[DIMS1]]#1 : index) stride(%[[DIMS0]]#2 : index) start_idx(%[[C0]] : index) {stride_in_bytes = true}
-!HOST: %[[ADDROF:.*]] = fir.box_addr %[[ARG0_DECL]]#1 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
+!HOST: %[[ADDROF:.*]] = fir.box_addr %[[ARG0_DECL]]#0 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
!HOST: %[[MAP:.*]] = omp.map_info var_ptr(%[[ADDROF]] : !fir.ref<!fir.array<?xi32>>, !fir.array<?xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<?xi32>> {name = "arr_read_write(2:5)"}
!HOST: omp.target map_entries(%[[MAP]] -> %{{.*}}, {{.*}} -> {{.*}} : !fir.ref<!fir.array<?xi32>>, !fir.ref<i32>) {
subroutine assumed_shape_array(arr_read_write)
@@ -68,13 +68,12 @@ end subroutine assumed_shape_array
!HOST: %[[ARG0_SHAPE:.*]] = fir.shape %{{.*}} : (index) -> !fir.shape<1>
!HOST: %[[ARG0_DECL:.*]]:2 = hlfir.declare %[[ARG0]](%[[ARG0_SHAPE]]) {fortran_attrs = #fir.var_attrs<intent_inout>, uniq_name = "_QMassumed_array_routinesFassumed_size_arrayEarr_read_write"} : (!fir.ref<!fir.array<?xi32>>, !fir.shape<1>) -> (!fir.box<!fir.array<?xi32>>, !fir.ref<!fir.array<?xi32>>)
!HOST: %[[ALLOCA:.*]] = fir.alloca i32 {bindc_name = "i", uniq_name = "_QMassumed_array_routinesFassumed_size_arrayEi"}
-!HOST: %[[C0:.*]] = arith.constant 1 : index
-!HOST: %[[C1:.*]] = arith.constant 1 : index
-!HOST: %[[C2:.*]] = arith.constant 4 : index
-!HOST: %[[DIFF:.*]] = arith.subi %[[C2]], %[[C1]] : index
-!HOST: %[[EXT:.*]] = arith.addi %[[DIFF]], %[[C0]] : index
-!HOST: %[[BOUNDS:.*]] = omp.bounds lower_bound(%[[C1]] : index) upper_bound(%[[C2]] : index) extent(%[[EXT]] : index) stride(%[[C0]] : index) start_idx(%[[C0]] : index)
-!HOST: %[[MAP:.*]] = omp.map_info var_ptr(%[[ARG0_DECL]]#1 : !fir.ref<!fir.array<?xi32>>, !fir.array<?xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<?xi32>> {name = "arr_read_write(2:5)"}
+!HOST: %[[DIMS0:.*]]:3 = fir.box_dims %[[ARG0_DECL]]#0, %c0{{.*}} : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index)
+!HOST: %[[C4_1:.*]] = arith.subi %c4, %c1{{.*}} : index
+!HOST: %[[EXT:.*]] = arith.addi %[[C4_1]], %c1{{.*}} : index
+!HOST: %[[BOUNDS:.*]] = omp.bounds lower_bound(%c1{{.*}} : index) upper_bound(%c4{{.*}} : index) extent(%[[EXT]] : index) stride(%[[DIMS0]]#2 : index) start_idx(%c1{{.*}} : index) {stride_in_bytes = true}
+!HOST: %[[ADDR:.*]] = fir.box_addr %[[ARG0_DECL]]#0 : (!fir.box<!fir.array<?xi32>>) -> !fir.ref<!fir.array<?xi32>>
+!HOST: %[[MAP:.*]] = omp.map_info var_ptr(%[[ADDR]] : !fir.ref<!fir.array<?xi32>>, !fir.array<?xi32>) map_clauses(tofrom) capture(ByRef) bounds(%7) -> !fir.ref<!fir.array<?xi32>> {name = "arr_read_write(2:5)"}
!HOST: omp.target map_entries(%[[MAP]] -> %{{.*}}, {{.*}} -> {{.*}} : !fir.ref<!fir.array<?xi32>>, !fir.ref<i32>) {
subroutine assumed_size_array(arr_read_write)
integer, intent(inout) :: arr_read_write(*)
diff --git a/flang/test/Lower/OpenMP/target.f90 b/flang/test/Lower/OpenMP/target.f90
index 5ca3a08d8a8b2..e9255cc9b1c0f 100644
--- a/flang/test/Lower/OpenMP/target.f90
+++ b/flang/test/Lower/OpenMP/target.f90
@@ -145,7 +145,7 @@ subroutine omp_target_update_to
!CHECK-DAG: %[[A_DECL:.*]]:2 = hlfir.declare %{{.*}}(%{{.*}})
!CHECK-DAG: %[[BOUNDS:.*]] = omp.bounds
- !CHECK: %[[TO_MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>)
+ !CHECK: %[[TO_MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>)
!CHECK-SAME: map_clauses(to) capture(ByRef)
!CHECK-SAME: bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
@@ -164,7 +164,7 @@ subroutine omp_target_update_from
!CHECK-DAG: %[[A_DECL:.*]]:2 = hlfir.declare %{{.*}}(%{{.*}})
!CHECK-DAG: %[[BOUNDS:.*]] = omp.bounds
- !CHECK: %[[FROM_MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>)
+ !CHECK: %[[FROM_MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>)
!CHECK-SAME: map_clauses(from) capture(ByRef)
!CHECK-SAME: bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
@@ -232,7 +232,7 @@ subroutine omp_target_data
!CHECK: %[[A_DECL:.*]]:2 = hlfir.declare %[[VAL_0]](%{{.*}}) {uniq_name = "_QFomp_target_dataEa"} : (!fir.ref<!fir.array<1024xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<1024xi32>>, !fir.ref<!fir.array<1024xi32>>)
integer :: a(1024)
!CHECK: %[[BOUNDS:.*]] = omp.bounds lower_bound({{.*}}) upper_bound({{.*}}) extent({{.*}}) stride({{.*}}) start_idx({{.*}})
- !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
+ !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[A_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
!CHECK: omp.target_data map_entries(%[[MAP]] : !fir.ref<!fir.array<1024xi32>>) {
!$omp target data map(tofrom: a)
!CHECK: %[[C10:.*]] = arith.constant 10 : i32
@@ -254,14 +254,14 @@ subroutine omp_target_data_mt
!CHECK: %[[VAR_B:.*]] = fir.alloca !fir.array<1024xi32> {bindc_name = "b", uniq_name = "_QFomp_target_data_mtEb"}
!CHECK: %[[VAR_B_DECL:.*]]:2 = hlfir.declare %[[VAR_B]](%{{.*}}) {uniq_name = "_QFomp_target_data_mtEb"} : (!fir.ref<!fir.array<1024xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<1024xi32>>, !fir.ref<!fir.array<1024xi32>>)
!CHECK: %[[BOUNDS_A:.*]] = omp.bounds lower_bound({{.*}}) upper_bound({{.*}}) extent({{.*}}) stride({{.*}}) start_idx({{.*}})
- !CHECK: %[[MAP_A:.*]] = omp.map_info var_ptr(%[[VAR_A_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS_A]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
+ !CHECK: %[[MAP_A:.*]] = omp.map_info var_ptr(%[[VAR_A_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS_A]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
!CHECK: omp.target_data map_entries(%[[MAP_A]] : !fir.ref<!fir.array<1024xi32>>) {
!$omp target data map(a)
!CHECK: omp.terminator
!$omp end target data
!CHECK: }
!CHECK: %[[BOUNDS_B:.*]] = omp.bounds lower_bound({{.*}}) upper_bound({{.*}}) extent({{.*}}) stride({{.*}}) start_idx({{.*}})
- !CHECK: %[[MAP_B:.*]] = omp.map_info var_ptr(%[[VAR_B_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(always, from) capture(ByRef) bounds(%[[BOUNDS_B]]) -> !fir.ref<!fir.array<1024xi32>> {name = "b"}
+ !CHECK: %[[MAP_B:.*]] = omp.map_info var_ptr(%[[VAR_B_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(always, from) capture(ByRef) bounds(%[[BOUNDS_B]]) -> !fir.ref<!fir.array<1024xi32>> {name = "b"}
!CHECK: omp.target_data map_entries(%[[MAP_B]] : !fir.ref<!fir.array<1024xi32>>) {
!$omp target data map(always, from : b)
!CHECK: omp.terminator
@@ -278,7 +278,7 @@ subroutine omp_target
!CHECK: %[[VAL_1:.*]]:2 = hlfir.declare %{{.*}}(%{{.*}}) {uniq_name = "_QFomp_targetEa"} : (!fir.ref<!fir.array<1024xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<1024xi32>>, !fir.ref<!fir.array<1024xi32>>)
integer :: a(1024)
!CHECK: %[[BOUNDS:.*]] = omp.bounds lower_bound({{.*}}) upper_bound({{.*}}) extent({{.*}}) stride({{.*}}) start_idx({{.*}})
- !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[VAL_1]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
+ !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[VAL_1]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
!CHECK: omp.target map_entries(%[[MAP]] -> %[[ARG_0:.*]] : !fir.ref<!fir.array<1024xi32>>) {
!CHECK: ^bb0(%[[ARG_0]]: !fir.ref<!fir.array<1024xi32>>):
!$omp target map(tofrom: a)
@@ -369,12 +369,12 @@ subroutine omp_target_implicit_bounds(n)
!CHECK: %[[VAL_8:.*]] = fir.alloca !fir.array<?xi32>, %[[VAL_7]] {bindc_name = "a", uniq_name = "_QFomp_target_implicit_boundsEa"}
!CHECK: %[[VAL_9:.*]] = fir.shape %[[VAL_7]] : (index) -> !fir.shape<1>
!CHECK: %[[VAL_10:.*]]:2 = hlfir.declare %[[VAL_8]](%[[VAL_9]]) {uniq_name = "_QFomp_target_implicit_boundsEa"} : (!fir.ref<!fir.array<?xi32>>, !fir.shape<1>) -> (!fir.box<!fir.array<?xi32>>, !fir.ref<!fir.array<?xi32>>)
- !CHECK: %[[VAL_11:.*]] = arith.constant 1 : index
- !CHECK: %[[VAL_12:.*]] = arith.constant 0 : index
- !CHECK: %[[VAL_13:.*]] = arith.subi %[[VAL_7]], %[[VAL_11]] : index
+ !CHECK: %[[DIMS0:.*]]:3 = fir.box_dims %9#0, %c0{{.*}} : (!fir.box<!fir.array<?xi32>>, index) -> (index, index, index)
+ !CHECK: %[[UB:.*]] = arith.subi %[[DIMS0]]#1, %c1{{.*}} : index
+
integer :: n
integer :: a(n)
- !CHECK: %[[VAL_14:.*]] = omp.bounds lower_bound(%[[VAL_12]] : index) upper_bound(%[[VAL_13]] : index) extent(%[[VAL_7]] : index) stride(%[[VAL_11]] : index) start_idx(%[[VAL_11]] : index)
+ !CHECK: %[[VAL_14:.*]] = omp.bounds lower_bound(%c0{{.*}} : index) upper_bound(%[[UB]] : index) extent(%[[DIMS0]]#1 : index) stride(%[[DIMS0]]#2 : index) start_idx(%c1{{.*}} : index) {stride_in_bytes = true}
!CHECK: %[[VAL_15:.*]] = omp.map_info var_ptr(%[[VAL_10]]#1 : !fir.ref<!fir.array<?xi32>>, !fir.array<?xi32>) map_clauses(implicit, tofrom) capture(ByRef) bounds(%[[VAL_14]]) -> !fir.ref<!fir.array<?xi32>> {name = "a"}
!CHECK: %[[VAL_16:.*]] = omp.map_info var_ptr(%[[VAL_COPY]] : !fir.ref<i32>, i32) map_clauses(implicit, exit_release_or_enter_alloc) capture(ByCopy) -> !fir.ref<i32> {name = ""}
!CHECK: omp.target map_entries(%[[VAL_15]] -> %[[VAL_17:.*]], %[[VAL_16]] -> %[[VAL_18:.*]] : !fir.ref<!fir.array<?xi32>>, !fir.ref<i32>) {
@@ -500,7 +500,7 @@ subroutine omp_target_parallel_do
!CHECK: %[[C0:.*]] = arith.constant 0 : index
!CHECK: %[[SUB:.*]] = arith.subi %[[C1024]], %[[C1]] : index
!CHECK: %[[BOUNDS:.*]] = omp.bounds lower_bound(%[[C0]] : index) upper_bound(%[[SUB]] : index) extent(%[[C1024]] : index) stride(%[[C1]] : index) start_idx(%[[C1]] : index)
- !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[VAL_0_DECL]]#1 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
+ !CHECK: %[[MAP:.*]] = omp.map_info var_ptr(%[[VAL_0_DECL]]#0 : !fir.ref<!fir.array<1024xi32>>, !fir.array<1024xi32>) map_clauses(tofrom) capture(ByRef) bounds(%[[BOUNDS]]) -> !fir.ref<!fir.array<1024xi32>> {name = "a"}
!CHECK: omp.target map_entries(%[[MAP]] -> %[[ARG_0:.*]], %{{.*}} -> %{{.*}} : !fir.ref<!fir.array<1024xi32>>, !fir.ref<i32>) {
!CHECK: ^bb0(%[[ARG_0]]: !fir.ref<!fir.array<1024xi32>>, %{{.*}}: !fir.ref<i32>):
!CHECK: %[[VAL_0_DECL:.*]]:2 = hlfir.declare %[[ARG_0]](%{{.*}}) {uniq_name = "_QFomp_target_parallel_doEa"} : (!fir.ref<!fir.array<1024xi32>>, !fir.shape<1>) -> (!fir.ref<!fir.array<1024xi32>>, !fir.ref<!fir.array<1024xi32>>)
>From 243bfed68367263cfc3fb3f396660acf60051fbf Mon Sep 17 00:00:00 2001
From: Artem Dergachev <adergachev at apple.com>
Date: Thu, 1 Feb 2024 13:07:21 -0800
Subject: [PATCH 05/12] [analyzer][HTMLRewriter] Cache partial rewrite results.
(#80220)
This is a follow-up for 721dd3bc2 [analyzer] NFC: Don't regenerate
duplicate HTML reports.
Because HTMLRewriter re-runs the Lexer for syntax highlighting and macro
expansion purposes, it may get fairly expensive when the rewriter is
invoked multiple times on the same file. In the static analyzer (which
uses HTMLRewriter for HTML output mode) we only get away with this
because there are usually very few reports emitted per file. But if loud
checkers are enabled, such as `webkit.*`, this may explode in complexity
and even cause the compiler to run over the 32-bit SourceLocation
addressing limit.
This patch caches intermediate results so that re-lexing only needed to
happen once.
As the clever __COUNTER__ test demonstrates, "once" is still too many.
Ideally we shouldn't re-lex anything at all, which remains a TODO.
---
.../include/clang/Rewrite/Core/HTMLRewrite.h | 15 +-
clang/lib/Rewrite/HTMLRewrite.cpp | 135 +++++++++++++++---
.../StaticAnalyzer/Core/HTMLDiagnostics.cpp | 10 +-
.../test/Analysis/html_diagnostics/counter.c | 29 ++++
4 files changed, 162 insertions(+), 27 deletions(-)
create mode 100644 clang/test/Analysis/html_diagnostics/counter.c
diff --git a/clang/include/clang/Rewrite/Core/HTMLRewrite.h b/clang/include/clang/Rewrite/Core/HTMLRewrite.h
index 340411e553479..eecf589632746 100644
--- a/clang/include/clang/Rewrite/Core/HTMLRewrite.h
+++ b/clang/include/clang/Rewrite/Core/HTMLRewrite.h
@@ -24,6 +24,15 @@ class RewriteBuffer;
class Preprocessor;
namespace html {
+ struct RelexRewriteCache;
+ using RelexRewriteCacheRef = std::shared_ptr<RelexRewriteCache>;
+
+ /// If you need to rewrite the same file multiple times, you can instantiate
+ /// a RelexRewriteCache and refer functions such as SyntaxHighlight()
+ /// and HighlightMacros() to it so that to avoid re-lexing the file each time.
+ /// The cache may outlive the rewriter as long as cached FileIDs and source
+ /// locations continue to make sense for the translation unit as a whole.
+ RelexRewriteCacheRef instantiateRelexRewriteCache();
/// HighlightRange - Highlight a range in the source code with the specified
/// start/end tags. B/E must be in the same file. This ensures that
@@ -67,13 +76,15 @@ namespace html {
/// SyntaxHighlight - Relex the specified FileID and annotate the HTML with
/// information about keywords, comments, etc.
- void SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP);
+ void SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP,
+ RelexRewriteCacheRef Cache = nullptr);
/// HighlightMacros - This uses the macro table state from the end of the
/// file, to reexpand macros and insert (into the HTML) information about the
/// macro expansions. This won't be perfectly perfect, but it will be
/// reasonably close.
- void HighlightMacros(Rewriter &R, FileID FID, const Preprocessor &PP);
+ void HighlightMacros(Rewriter &R, FileID FID, const Preprocessor &PP,
+ RelexRewriteCacheRef Cache = nullptr);
} // end html namespace
} // end clang namespace
diff --git a/clang/lib/Rewrite/HTMLRewrite.cpp b/clang/lib/Rewrite/HTMLRewrite.cpp
index 083a9c09297e1..a96ca0764ae73 100644
--- a/clang/lib/Rewrite/HTMLRewrite.cpp
+++ b/clang/lib/Rewrite/HTMLRewrite.cpp
@@ -21,8 +21,10 @@
#include "llvm/Support/MemoryBuffer.h"
#include "llvm/Support/raw_ostream.h"
#include <memory>
-using namespace clang;
+using namespace clang;
+using namespace llvm;
+using namespace html;
/// HighlightRange - Highlight a range in the source code with the specified
/// start/end tags. B/E must be in the same file. This ensures that
@@ -104,6 +106,32 @@ void html::HighlightRange(RewriteBuffer &RB, unsigned B, unsigned E,
}
}
+namespace clang::html {
+struct RelexRewriteCache {
+ // These structs mimic input arguments of HighlightRange().
+ struct Highlight {
+ SourceLocation B, E;
+ std::string StartTag, EndTag;
+ bool IsTokenRange;
+ };
+ struct RawHighlight {
+ unsigned B, E;
+ std::string StartTag, EndTag;
+ };
+
+ // SmallVector isn't appropriate because these vectors are almost never small.
+ using HighlightList = std::vector<Highlight>;
+ using RawHighlightList = std::vector<RawHighlight>;
+
+ DenseMap<FileID, RawHighlightList> SyntaxHighlights;
+ DenseMap<FileID, HighlightList> MacroHighlights;
+};
+} // namespace clang::html
+
+html::RelexRewriteCacheRef html::instantiateRelexRewriteCache() {
+ return std::make_shared<RelexRewriteCache>();
+}
+
void html::EscapeText(Rewriter &R, FileID FID,
bool EscapeSpaces, bool ReplaceTabs) {
@@ -442,13 +470,18 @@ input.spoilerhider:checked + label + .spoiler{
/// information about keywords, macro expansions etc. This uses the macro
/// table state from the end of the file, so it won't be perfectly perfect,
/// but it will be reasonably close.
-void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP) {
- RewriteBuffer &RB = R.getEditBuffer(FID);
+static void SyntaxHighlightImpl(
+ Rewriter &R, FileID FID, const Preprocessor &PP,
+ llvm::function_ref<void(RewriteBuffer &, unsigned, unsigned, const char *,
+ const char *, const char *)>
+ HighlightRangeCallback) {
+ RewriteBuffer &RB = R.getEditBuffer(FID);
const SourceManager &SM = PP.getSourceManager();
llvm::MemoryBufferRef FromFile = SM.getBufferOrFake(FID);
+ const char *BufferStart = FromFile.getBuffer().data();
+
Lexer L(FID, FromFile, SM, PP.getLangOpts());
- const char *BufferStart = L.getBuffer().data();
// Inform the preprocessor that we want to retain comments as tokens, so we
// can highlight them.
@@ -475,13 +508,13 @@ void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP) {
// If this is a pp-identifier, for a keyword, highlight it as such.
if (Tok.isNot(tok::identifier))
- HighlightRange(RB, TokOffs, TokOffs+TokLen, BufferStart,
- "<span class='keyword'>", "</span>");
+ HighlightRangeCallback(RB, TokOffs, TokOffs + TokLen, BufferStart,
+ "<span class='keyword'>", "</span>");
break;
}
case tok::comment:
- HighlightRange(RB, TokOffs, TokOffs+TokLen, BufferStart,
- "<span class='comment'>", "</span>");
+ HighlightRangeCallback(RB, TokOffs, TokOffs + TokLen, BufferStart,
+ "<span class='comment'>", "</span>");
break;
case tok::utf8_string_literal:
// Chop off the u part of u8 prefix
@@ -498,8 +531,8 @@ void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP) {
[[fallthrough]];
case tok::string_literal:
// FIXME: Exclude the optional ud-suffix from the highlighted range.
- HighlightRange(RB, TokOffs, TokOffs+TokLen, BufferStart,
- "<span class='string_literal'>", "</span>");
+ HighlightRangeCallback(RB, TokOffs, TokOffs + TokLen, BufferStart,
+ "<span class='string_literal'>", "</span>");
break;
case tok::hash: {
// If this is a preprocessor directive, all tokens to end of line are too.
@@ -516,8 +549,8 @@ void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP) {
}
// Find end of line. This is a hack.
- HighlightRange(RB, TokOffs, TokEnd, BufferStart,
- "<span class='directive'>", "</span>");
+ HighlightRangeCallback(RB, TokOffs, TokEnd, BufferStart,
+ "<span class='directive'>", "</span>");
// Don't skip the next token.
continue;
@@ -527,12 +560,43 @@ void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP) {
L.LexFromRawLexer(Tok);
}
}
+void html::SyntaxHighlight(Rewriter &R, FileID FID, const Preprocessor &PP,
+ RelexRewriteCacheRef Cache) {
+ RewriteBuffer &RB = R.getEditBuffer(FID);
+ const SourceManager &SM = PP.getSourceManager();
+ llvm::MemoryBufferRef FromFile = SM.getBufferOrFake(FID);
+ const char *BufferStart = FromFile.getBuffer().data();
+
+ if (Cache) {
+ auto CacheIt = Cache->SyntaxHighlights.find(FID);
+ if (CacheIt != Cache->SyntaxHighlights.end()) {
+ for (const RelexRewriteCache::RawHighlight &H : CacheIt->second) {
+ HighlightRange(RB, H.B, H.E, BufferStart, H.StartTag.data(),
+ H.EndTag.data());
+ }
+ return;
+ }
+ }
+
+ // "Every time you would call HighlightRange, cache the inputs as well."
+ auto HighlightRangeCallback = [&](RewriteBuffer &RB, unsigned B, unsigned E,
+ const char *BufferStart,
+ const char *StartTag, const char *EndTag) {
+ HighlightRange(RB, B, E, BufferStart, StartTag, EndTag);
+
+ if (Cache)
+ Cache->SyntaxHighlights[FID].push_back({B, E, StartTag, EndTag});
+ };
+
+ SyntaxHighlightImpl(R, FID, PP, HighlightRangeCallback);
+}
+
+static void HighlightMacrosImpl(
+ Rewriter &R, FileID FID, const Preprocessor &PP,
+ llvm::function_ref<void(Rewriter &, SourceLocation, SourceLocation,
+ const char *, const char *, bool)>
+ HighlightRangeCallback) {
-/// HighlightMacros - This uses the macro table state from the end of the
-/// file, to re-expand macros and insert (into the HTML) information about the
-/// macro expansions. This won't be perfectly perfect, but it will be
-/// reasonably close.
-void html::HighlightMacros(Rewriter &R, FileID FID, const Preprocessor& PP) {
// Re-lex the raw token stream into a token buffer.
const SourceManager &SM = PP.getSourceManager();
std::vector<Token> TokenStream;
@@ -659,11 +723,44 @@ void html::HighlightMacros(Rewriter &R, FileID FID, const Preprocessor& PP) {
// get highlighted.
Expansion = "<span class='macro_popup'>" + Expansion + "</span></span>";
- HighlightRange(R, LLoc.getBegin(), LLoc.getEnd(), "<span class='macro'>",
- Expansion.c_str(), LLoc.isTokenRange());
+ HighlightRangeCallback(R, LLoc.getBegin(), LLoc.getEnd(),
+ "<span class='macro'>", Expansion.c_str(),
+ LLoc.isTokenRange());
}
// Restore the preprocessor's old state.
TmpPP.setDiagnostics(*OldDiags);
TmpPP.setPragmasEnabled(PragmasPreviouslyEnabled);
}
+
+/// HighlightMacros - This uses the macro table state from the end of the
+/// file, to re-expand macros and insert (into the HTML) information about the
+/// macro expansions. This won't be perfectly perfect, but it will be
+/// reasonably close.
+void html::HighlightMacros(Rewriter &R, FileID FID, const Preprocessor &PP,
+ RelexRewriteCacheRef Cache) {
+ if (Cache) {
+ auto CacheIt = Cache->MacroHighlights.find(FID);
+ if (CacheIt != Cache->MacroHighlights.end()) {
+ for (const RelexRewriteCache::Highlight &H : CacheIt->second) {
+ HighlightRange(R, H.B, H.E, H.StartTag.data(), H.EndTag.data(),
+ H.IsTokenRange);
+ }
+ return;
+ }
+ }
+
+ // "Every time you would call HighlightRange, cache the inputs as well."
+ auto HighlightRangeCallback = [&](Rewriter &R, SourceLocation B,
+ SourceLocation E, const char *StartTag,
+ const char *EndTag, bool isTokenRange) {
+ HighlightRange(R, B, E, StartTag, EndTag, isTokenRange);
+
+ if (Cache) {
+ Cache->MacroHighlights[FID].push_back(
+ {B, E, StartTag, EndTag, isTokenRange});
+ }
+ };
+
+ HighlightMacrosImpl(R, FID, PP, HighlightRangeCallback);
+}
diff --git a/clang/lib/StaticAnalyzer/Core/HTMLDiagnostics.cpp b/clang/lib/StaticAnalyzer/Core/HTMLDiagnostics.cpp
index 86947b7929e9b..fb5030d373c2f 100644
--- a/clang/lib/StaticAnalyzer/Core/HTMLDiagnostics.cpp
+++ b/clang/lib/StaticAnalyzer/Core/HTMLDiagnostics.cpp
@@ -69,6 +69,8 @@ class HTMLDiagnostics : public PathDiagnosticConsumer {
const Preprocessor &PP;
const bool SupportsCrossFileDiagnostics;
llvm::StringSet<> EmittedHashes;
+ html::RelexRewriteCacheRef RewriterCache =
+ html::instantiateRelexRewriteCache();
public:
HTMLDiagnostics(PathDiagnosticConsumerOptions DiagOpts,
@@ -309,10 +311,6 @@ void HTMLDiagnostics::ReportDiag(const PathDiagnostic& D,
return;
}
- // FIXME: This causes each file to be re-parsed and syntax-highlighted
- // and macro-expanded separately for each report. We could cache such rewrites
- // across all reports and only re-do the part that's actually different:
- // the warning/note bubbles.
std::string report = GenerateHTML(D, R, SMgr, path, declName.c_str());
if (report.empty()) {
llvm::errs() << "warning: no diagnostics generated for main file.\n";
@@ -882,8 +880,8 @@ void HTMLDiagnostics::RewriteFile(Rewriter &R, const PathPieces &path,
// If we have a preprocessor, relex the file and syntax highlight.
// We might not have a preprocessor if we come from a deserialized AST file,
// for example.
- html::SyntaxHighlight(R, FID, PP);
- html::HighlightMacros(R, FID, PP);
+ html::SyntaxHighlight(R, FID, PP, RewriterCache);
+ html::HighlightMacros(R, FID, PP, RewriterCache);
}
void HTMLDiagnostics::HandlePiece(Rewriter &R, FileID BugFileID,
diff --git a/clang/test/Analysis/html_diagnostics/counter.c b/clang/test/Analysis/html_diagnostics/counter.c
new file mode 100644
index 0000000000000..22f722424372f
--- /dev/null
+++ b/clang/test/Analysis/html_diagnostics/counter.c
@@ -0,0 +1,29 @@
+// RUN: rm -fR %t
+// RUN: mkdir %t
+// RUN: %clang_analyze_cc1 -analyzer-checker=core \
+// RUN: -analyzer-output=html -o %t -verify %s
+// RUN: grep -v CHECK %t/report-*.html | FileCheck %s
+
+
+void foo() {
+ int *x = 0;
+ *x = __COUNTER__; // expected-warning{{Dereference of null pointer (loaded from variable 'x')}}
+}
+
+void bar() {
+ int *y;
+ *y = __COUNTER__; // expected-warning{{Dereference of undefined pointer value (loaded from variable 'y')}}
+}
+
+// The checks below confirm that both reports have the same values for __COUNTER__.
+//
+// FIXME: The correct values are (0, 1, 0, 1). Because we re-lex the file in order
+// to detect macro expansions for HTML report purposes, they turn into (2, 3, 2, 3)
+// by the time we emit HTML. But at least it's better than (2, 3, 4, 5)
+// which would have been the case if we re-lexed the file *each* time we
+// emitted an HTML report.
+
+// CHECK: <span class='macro'>__COUNTER__<span class='macro_popup'>2</span></span>
+// CHECK: <span class='macro'>__COUNTER__<span class='macro_popup'>3</span></span>
+// CHECK: <span class='macro'>__COUNTER__<span class='macro_popup'>2</span></span>
+// CHECK: <span class='macro'>__COUNTER__<span class='macro_popup'>3</span></span>
>From 8ad14b6d90121d2d0687a3a7f6f6c6f2b34c4aa7 Mon Sep 17 00:00:00 2001
From: Alexey Bataev <5361294+alexey-bataev at users.noreply.github.com>
Date: Thu, 1 Feb 2024 16:07:38 -0500
Subject: [PATCH 06/12] [TTI]Add support for strided loads/stores.
Added basic legality check and cost estimation functions for strided loads and stores.
These interfaces will be built upon in https://github.com/llvm/llvm-project/pull/80310.
Reviewers: preames
Reviewed By: preames
Pull Request: https://github.com/llvm/llvm-project/pull/80329
---
.../llvm/Analysis/TargetTransformInfo.h | 34 +++++++++++++++++++
.../llvm/Analysis/TargetTransformInfoImpl.h | 12 +++++++
llvm/lib/Analysis/TargetTransformInfo.cpp | 15 ++++++++
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 23 +++++++++++++
.../Target/RISCV/RISCVTargetTransformInfo.h | 11 ++++++
5 files changed, 95 insertions(+)
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfo.h b/llvm/include/llvm/Analysis/TargetTransformInfo.h
index 3b615bc700bbb..58577a6b6eb5c 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfo.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfo.h
@@ -781,6 +781,9 @@ class TargetTransformInfo {
/// Return true if the target supports masked expand load.
bool isLegalMaskedExpandLoad(Type *DataType) const;
+ /// Return true if the target supports strided load.
+ bool isLegalStridedLoadStore(Type *DataType, Align Alignment) const;
+
/// Return true if this is an alternating opcode pattern that can be lowered
/// to a single instruction on the target. In X86 this is for the addsub
/// instruction which corrsponds to a Shuffle + Fadd + FSub pattern in IR.
@@ -1412,6 +1415,20 @@ class TargetTransformInfo {
Align Alignment, TTI::TargetCostKind CostKind = TTI::TCK_RecipThroughput,
const Instruction *I = nullptr) const;
+ /// \return The cost of strided memory operations.
+ /// \p Opcode - is a type of memory access Load or Store
+ /// \p DataTy - a vector type of the data to be loaded or stored
+ /// \p Ptr - pointer [or vector of pointers] - address[es] in memory
+ /// \p VariableMask - true when the memory access is predicated with a mask
+ /// that is not a compile-time constant
+ /// \p Alignment - alignment of single element
+ /// \p I - the optional original context instruction, if one exists, e.g. the
+ /// load/store to transform or the call to the gather/scatter intrinsic
+ InstructionCost getStridedMemoryOpCost(
+ unsigned Opcode, Type *DataTy, const Value *Ptr, bool VariableMask,
+ Align Alignment, TTI::TargetCostKind CostKind = TTI::TCK_RecipThroughput,
+ const Instruction *I = nullptr) const;
+
/// \return The cost of the interleaved memory operation.
/// \p Opcode is the memory operation code
/// \p VecTy is the vector type of the interleaved access.
@@ -1848,6 +1865,7 @@ class TargetTransformInfo::Concept {
Align Alignment) = 0;
virtual bool isLegalMaskedCompressStore(Type *DataType) = 0;
virtual bool isLegalMaskedExpandLoad(Type *DataType) = 0;
+ virtual bool isLegalStridedLoadStore(Type *DataType, Align Alignment) = 0;
virtual bool isLegalAltInstr(VectorType *VecTy, unsigned Opcode0,
unsigned Opcode1,
const SmallBitVector &OpcodeMask) const = 0;
@@ -2023,6 +2041,11 @@ class TargetTransformInfo::Concept {
bool VariableMask, Align Alignment,
TTI::TargetCostKind CostKind,
const Instruction *I = nullptr) = 0;
+ virtual InstructionCost
+ getStridedMemoryOpCost(unsigned Opcode, Type *DataTy, const Value *Ptr,
+ bool VariableMask, Align Alignment,
+ TTI::TargetCostKind CostKind,
+ const Instruction *I = nullptr) = 0;
virtual InstructionCost getInterleavedMemoryOpCost(
unsigned Opcode, Type *VecTy, unsigned Factor, ArrayRef<unsigned> Indices,
@@ -2341,6 +2364,9 @@ class TargetTransformInfo::Model final : public TargetTransformInfo::Concept {
bool isLegalMaskedExpandLoad(Type *DataType) override {
return Impl.isLegalMaskedExpandLoad(DataType);
}
+ bool isLegalStridedLoadStore(Type *DataType, Align Alignment) override {
+ return Impl.isLegalStridedLoadStore(DataType, Alignment);
+ }
bool isLegalAltInstr(VectorType *VecTy, unsigned Opcode0, unsigned Opcode1,
const SmallBitVector &OpcodeMask) const override {
return Impl.isLegalAltInstr(VecTy, Opcode0, Opcode1, OpcodeMask);
@@ -2671,6 +2697,14 @@ class TargetTransformInfo::Model final : public TargetTransformInfo::Concept {
return Impl.getGatherScatterOpCost(Opcode, DataTy, Ptr, VariableMask,
Alignment, CostKind, I);
}
+ InstructionCost
+ getStridedMemoryOpCost(unsigned Opcode, Type *DataTy, const Value *Ptr,
+ bool VariableMask, Align Alignment,
+ TTI::TargetCostKind CostKind,
+ const Instruction *I = nullptr) override {
+ return Impl.getStridedMemoryOpCost(Opcode, DataTy, Ptr, VariableMask,
+ Alignment, CostKind, I);
+ }
InstructionCost getInterleavedMemoryOpCost(
unsigned Opcode, Type *VecTy, unsigned Factor, ArrayRef<unsigned> Indices,
Align Alignment, unsigned AddressSpace, TTI::TargetCostKind CostKind,
diff --git a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
index 9958b4daa6ed8..3d5db96e86b80 100644
--- a/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
+++ b/llvm/include/llvm/Analysis/TargetTransformInfoImpl.h
@@ -304,6 +304,10 @@ class TargetTransformInfoImplBase {
bool isLegalMaskedExpandLoad(Type *DataType) const { return false; }
+ bool isLegalStridedLoadStore(Type *DataType, Align Alignment) const {
+ return false;
+ }
+
bool enableOrderedReductions() const { return false; }
bool hasDivRemOp(Type *DataType, bool IsSigned) const { return false; }
@@ -687,6 +691,14 @@ class TargetTransformInfoImplBase {
return 1;
}
+ InstructionCost getStridedMemoryOpCost(unsigned Opcode, Type *DataTy,
+ const Value *Ptr, bool VariableMask,
+ Align Alignment,
+ TTI::TargetCostKind CostKind,
+ const Instruction *I = nullptr) const {
+ return InstructionCost::getInvalid();
+ }
+
unsigned getInterleavedMemoryOpCost(
unsigned Opcode, Type *VecTy, unsigned Factor, ArrayRef<unsigned> Indices,
Align Alignment, unsigned AddressSpace, TTI::TargetCostKind CostKind,
diff --git a/llvm/lib/Analysis/TargetTransformInfo.cpp b/llvm/lib/Analysis/TargetTransformInfo.cpp
index 8902dde37cbca..1f11f0d7dd620 100644
--- a/llvm/lib/Analysis/TargetTransformInfo.cpp
+++ b/llvm/lib/Analysis/TargetTransformInfo.cpp
@@ -500,6 +500,11 @@ bool TargetTransformInfo::isLegalMaskedExpandLoad(Type *DataType) const {
return TTIImpl->isLegalMaskedExpandLoad(DataType);
}
+bool TargetTransformInfo::isLegalStridedLoadStore(Type *DataType,
+ Align Alignment) const {
+ return TTIImpl->isLegalStridedLoadStore(DataType, Alignment);
+}
+
bool TargetTransformInfo::enableOrderedReductions() const {
return TTIImpl->enableOrderedReductions();
}
@@ -1037,6 +1042,16 @@ InstructionCost TargetTransformInfo::getGatherScatterOpCost(
Align Alignment, TTI::TargetCostKind CostKind, const Instruction *I) const {
InstructionCost Cost = TTIImpl->getGatherScatterOpCost(
Opcode, DataTy, Ptr, VariableMask, Alignment, CostKind, I);
+ assert((!Cost.isValid() || Cost >= 0) &&
+ "TTI should not produce negative costs!");
+ return Cost;
+}
+
+InstructionCost TargetTransformInfo::getStridedMemoryOpCost(
+ unsigned Opcode, Type *DataTy, const Value *Ptr, bool VariableMask,
+ Align Alignment, TTI::TargetCostKind CostKind, const Instruction *I) const {
+ InstructionCost Cost = TTIImpl->getStridedMemoryOpCost(
+ Opcode, DataTy, Ptr, VariableMask, Alignment, CostKind, I);
assert(Cost >= 0 && "TTI should not produce negative costs!");
return Cost;
}
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index fe1cdb2dfa423..cb48720cc1902 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -658,6 +658,29 @@ InstructionCost RISCVTTIImpl::getGatherScatterOpCost(
return NumLoads * MemOpCost;
}
+InstructionCost RISCVTTIImpl::getStridedMemoryOpCost(
+ unsigned Opcode, Type *DataTy, const Value *Ptr, bool VariableMask,
+ Align Alignment, TTI::TargetCostKind CostKind, const Instruction *I) {
+ if (((Opcode == Instruction::Load || Opcode == Instruction::Store) &&
+ !isLegalStridedLoadStore(DataTy, Alignment)) ||
+ (Opcode != Instruction::Load && Opcode != Instruction::Store))
+ return BaseT::getStridedMemoryOpCost(Opcode, DataTy, Ptr, VariableMask,
+ Alignment, CostKind, I);
+
+ if (CostKind == TTI::TCK_CodeSize)
+ return TTI::TCC_Basic;
+
+ // Cost is proportional to the number of memory operations implied. For
+ // scalable vectors, we use an estimate on that number since we don't
+ // know exactly what VL will be.
+ auto &VTy = *cast<VectorType>(DataTy);
+ InstructionCost MemOpCost =
+ getMemoryOpCost(Opcode, VTy.getElementType(), Alignment, 0, CostKind,
+ {TTI::OK_AnyValue, TTI::OP_None}, I);
+ unsigned NumLoads = getEstimatedVLFor(&VTy);
+ return NumLoads * MemOpCost;
+}
+
// Currently, these represent both throughput and codesize costs
// for the respective intrinsics. The costs in this table are simply
// instruction counts with the following adjustments made:
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
index 0747a778fe9a2..af36e9d5d5e88 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.h
@@ -143,6 +143,12 @@ class RISCVTTIImpl : public BasicTTIImplBase<RISCVTTIImpl> {
TTI::TargetCostKind CostKind,
const Instruction *I);
+ InstructionCost getStridedMemoryOpCost(unsigned Opcode, Type *DataTy,
+ const Value *Ptr, bool VariableMask,
+ Align Alignment,
+ TTI::TargetCostKind CostKind,
+ const Instruction *I);
+
InstructionCost getCastInstrCost(unsigned Opcode, Type *Dst, Type *Src,
TTI::CastContextHint CCH,
TTI::TargetCostKind CostKind,
@@ -250,6 +256,11 @@ class RISCVTTIImpl : public BasicTTIImplBase<RISCVTTIImpl> {
return ST->is64Bit() && !ST->hasVInstructionsI64();
}
+ bool isLegalStridedLoadStore(Type *DataType, Align Alignment) {
+ EVT DataTypeVT = TLI->getValueType(DL, DataType);
+ return TLI->isLegalStridedLoadStore(DataTypeVT, Alignment);
+ }
+
bool isVScaleKnownToBeAPowerOfTwo() const {
return TLI->isVScaleKnownToBeAPowerOfTwo();
}
>From 59e559067b1e47b72f3c631483c1f71919d5358e Mon Sep 17 00:00:00 2001
From: Philip Reames <preames at rivosinc.com>
Date: Thu, 1 Feb 2024 13:09:07 -0800
Subject: [PATCH 07/12] Revert "[RISCV] Refine cost on Min/Max reduction"
(#80340)
Reverts llvm/llvm-project#79402. Crash reported. On closer inspection,
this patch does not handle Intrinsic::maximum and Intrinsic::minimum.
---
.../Target/RISCV/RISCVTargetTransformInfo.cpp | 43 +-----
.../Analysis/CostModel/RISCV/reduce-max.ll | 140 +++++++++---------
.../Analysis/CostModel/RISCV/reduce-min.ll | 140 +++++++++---------
.../CostModel/RISCV/reduce-scalable-fp.ll | 84 +++++------
.../CostModel/RISCV/reduce-scalable-int.ll | 96 ++++++------
5 files changed, 237 insertions(+), 266 deletions(-)
diff --git a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
index cb48720cc1902..d1db47a6061e4 100644
--- a/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
+++ b/llvm/lib/Target/RISCV/RISCVTargetTransformInfo.cpp
@@ -970,42 +970,13 @@ RISCVTTIImpl::getMinMaxReductionCost(Intrinsic::ID IID, VectorType *Ty,
}
// IR Reduction is composed by two vmv and one rvv reduction instruction.
- unsigned SplitOp;
- SmallVector<unsigned, 3> Opcodes;
- switch (IID) {
- default:
- llvm_unreachable("Unsupported intrinsic");
- case Intrinsic::smax:
- SplitOp = RISCV::VMAX_VV;
- Opcodes = {RISCV::VMV_S_X, RISCV::VREDMAX_VS, RISCV::VMV_X_S};
- break;
- case Intrinsic::smin:
- SplitOp = RISCV::VMIN_VV;
- Opcodes = {RISCV::VMV_S_X, RISCV::VREDMIN_VS, RISCV::VMV_X_S};
- break;
- case Intrinsic::umax:
- SplitOp = RISCV::VMAXU_VV;
- Opcodes = {RISCV::VMV_S_X, RISCV::VREDMAXU_VS, RISCV::VMV_X_S};
- break;
- case Intrinsic::umin:
- SplitOp = RISCV::VMINU_VV;
- Opcodes = {RISCV::VMV_S_X, RISCV::VREDMINU_VS, RISCV::VMV_X_S};
- break;
- case Intrinsic::maxnum:
- SplitOp = RISCV::VFMAX_VV;
- Opcodes = {RISCV::VFMV_S_F, RISCV::VFREDMAX_VS, RISCV::VFMV_F_S};
- break;
- case Intrinsic::minnum:
- SplitOp = RISCV::VFMIN_VV;
- Opcodes = {RISCV::VFMV_S_F, RISCV::VFREDMIN_VS, RISCV::VFMV_F_S};
- break;
- }
- // Add a cost for data larger than LMUL8
- InstructionCost SplitCost =
- (LT.first > 1) ? (LT.first - 1) *
- getRISCVInstructionCost(SplitOp, LT.second, CostKind)
- : 0;
- return SplitCost + getRISCVInstructionCost(Opcodes, LT.second, CostKind);
+ InstructionCost BaseCost = 2;
+
+ if (CostKind == TTI::TCK_CodeSize)
+ return (LT.first - 1) + BaseCost;
+
+ unsigned VL = getEstimatedVLFor(Ty);
+ return (LT.first - 1) + BaseCost + Log2_32_Ceil(VL);
}
InstructionCost
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
index 533a77dbd9f71..f3436b54df90a 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-max.ll
@@ -51,14 +51,14 @@ define i32 @reduce_umax_i8(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umax_i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umax.v2i8(<2 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.umax.v4i8(<4 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.umax.v8i8(<8 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.umax.v16i8(<16 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.umax.v32i8(<32 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.umax.v64i8(<64 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.umax.v128i8(<128 x i8> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i8 @llvm.vector.reduce.umax.v1i8(<1 x i8> undef)
@@ -85,14 +85,14 @@ define i32 @reduce_umax_i16(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umax_i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umax.v2i16(<2 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.umax.v4i16(<4 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.umax.v8i16(<8 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.umax.v16i16(<16 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.umax.v32i16(<32 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.umax.v64i16(<64 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i16 @llvm.vector.reduce.umax.v128i16(<128 x i16> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i16 @llvm.vector.reduce.umax.v1i16(<1 x i16> undef)
@@ -115,18 +115,18 @@ define i32 @reduce_umax_i32(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umax_i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umax.v2i32(<2 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.umax.v4i32(<4 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.umax.v8i32(<8 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.umax.v16i32(<16 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.umax.v32i32(<32 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i32 @llvm.vector.reduce.umax.v64i32(<64 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i32 @llvm.vector.reduce.umax.v128i32(<128 x i32> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i32 @llvm.vector.reduce.umax.v1i32(<1 x i32> undef)
@@ -148,19 +148,19 @@ define i32 @reduce_umax_i64(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umax_i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umax.v2i64(<2 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.umax.v4i64(<4 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.umax.v8i64(<8 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.umax.v16i64(<16 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i64 @llvm.vector.reduce.umax.v32i64(<32 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i64 @llvm.vector.reduce.umax.v64i64(<64 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i64 @llvm.vector.reduce.umax.v128i64(<128 x i64> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i64 @llvm.vector.reduce.umax.v1i64(<1 x i64> undef)
@@ -221,14 +221,14 @@ define i32 @reduce_smax_i8(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smax_i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smax.v2i8(<2 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.smax.v4i8(<4 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.smax.v8i8(<8 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.smax.v16i8(<16 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.smax.v32i8(<32 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.smax.v64i8(<64 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.smax.v128i8(<128 x i8> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i8 @llvm.vector.reduce.smax.v1i8(<1 x i8> undef)
@@ -255,14 +255,14 @@ define i32 @reduce_smax_i16(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smax_i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smax.v2i16(<2 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.smax.v4i16(<4 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.smax.v8i16(<8 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.smax.v16i16(<16 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.smax.v32i16(<32 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.smax.v64i16(<64 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i16 @llvm.vector.reduce.smax.v128i16(<128 x i16> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i16 @llvm.vector.reduce.smax.v1i16(<1 x i16> undef)
@@ -285,18 +285,18 @@ define i32 @reduce_smax_i32(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smax_i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smax.v2i32(<2 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.smax.v4i32(<4 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.smax.v8i32(<8 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.smax.v16i32(<16 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.smax.v32i32(<32 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i32 @llvm.vector.reduce.smax.v64i32(<64 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i32 @llvm.vector.reduce.smax.v128i32(<128 x i32> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i32 @llvm.vector.reduce.smax.v1i32(<1 x i32> undef)
@@ -318,19 +318,19 @@ define i32 @reduce_smax_i64(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smax_i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smax.v2i64(<2 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.smax.v4i64(<4 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.smax.v8i64(<8 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.smax.v16i64(<16 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i64 @llvm.vector.reduce.smax.v32i64(<32 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i64 @llvm.vector.reduce.smax.v64i64(<64 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i64 @llvm.vector.reduce.smax.v128i64(<128 x i64> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i64 @llvm.vector.reduce.smax.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
index ae99118d4d3af..1964e35b81b1a 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-min.ll
@@ -51,14 +51,14 @@ define i32 @reduce_umin_i8(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umin_i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.umin.v2i8(<2 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.umin.v4i8(<4 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.umin.v8i8(<8 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.umin.v16i8(<16 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.umin.v32i8(<32 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.umin.v64i8(<64 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.umin.v128i8(<128 x i8> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i8 @llvm.vector.reduce.umin.v1i8(<1 x i8> undef)
@@ -85,14 +85,14 @@ define i32 @reduce_umin_i16(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umin_i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.umin.v2i16(<2 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.umin.v4i16(<4 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.umin.v8i16(<8 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.umin.v16i16(<16 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.umin.v32i16(<32 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.umin.v64i16(<64 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i16 @llvm.vector.reduce.umin.v128i16(<128 x i16> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i16 @llvm.vector.reduce.umin.v1i16(<1 x i16> undef)
@@ -115,18 +115,18 @@ define i32 @reduce_umin_i32(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umin_i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.umin.v2i32(<2 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.umin.v4i32(<4 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.umin.v8i32(<8 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.umin.v16i32(<16 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.umin.v32i32(<32 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i32 @llvm.vector.reduce.umin.v64i32(<64 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i32 @llvm.vector.reduce.umin.v128i32(<128 x i32> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i32 @llvm.vector.reduce.umin.v1i32(<1 x i32> undef)
@@ -148,19 +148,19 @@ define i32 @reduce_umin_i64(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_umin_i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.umin.v2i64(<2 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.umin.v4i64(<4 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.umin.v8i64(<8 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.umin.v16i64(<16 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i64 @llvm.vector.reduce.umin.v32i64(<32 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i64 @llvm.vector.reduce.umin.v64i64(<64 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i64 @llvm.vector.reduce.umin.v128i64(<128 x i64> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i64 @llvm.vector.reduce.umin.v1i64(<1 x i64> undef)
@@ -221,14 +221,14 @@ define i32 @reduce_smin_i8(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smin_i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i8 @llvm.vector.reduce.smin.v2i8(<2 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i8 @llvm.vector.reduce.smin.v4i8(<4 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i8 @llvm.vector.reduce.smin.v8i8(<8 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i8 @llvm.vector.reduce.smin.v16i8(<16 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i8 @llvm.vector.reduce.smin.v32i8(<32 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i8 @llvm.vector.reduce.smin.v64i8(<64 x i8> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i8 @llvm.vector.reduce.smin.v128i8(<128 x i8> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i8 @llvm.vector.reduce.smin.v1i8(<1 x i8> undef)
@@ -255,14 +255,14 @@ define i32 @reduce_smin_i16(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smin_i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i16 @llvm.vector.reduce.smin.v2i16(<2 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i16 @llvm.vector.reduce.smin.v4i16(<4 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i16 @llvm.vector.reduce.smin.v8i16(<8 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i16 @llvm.vector.reduce.smin.v16i16(<16 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i16 @llvm.vector.reduce.smin.v32i16(<32 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i16 @llvm.vector.reduce.smin.v64i16(<64 x i16> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V128 = call i16 @llvm.vector.reduce.smin.v128i16(<128 x i16> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i16 @llvm.vector.reduce.smin.v1i16(<1 x i16> undef)
@@ -285,18 +285,18 @@ define i32 @reduce_smin_i32(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smin_i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i32 @llvm.vector.reduce.smin.v2i32(<2 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i32 @llvm.vector.reduce.smin.v4i32(<4 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i32 @llvm.vector.reduce.smin.v8i32(<8 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i32 @llvm.vector.reduce.smin.v16i32(<16 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i32 @llvm.vector.reduce.smin.v32i32(<32 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V64 = call i32 @llvm.vector.reduce.smin.v64i32(<64 x i32> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V128 = call i32 @llvm.vector.reduce.smin.v128i32(<128 x i32> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i32 @llvm.vector.reduce.smin.v1i32(<1 x i32> undef)
@@ -318,19 +318,19 @@ define i32 @reduce_smin_i64(i32 %arg) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 7 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
-; CHECK-NEXT: Cost Model: Found an estimated cost of 31 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 12 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 undef
;
; SIZE-LABEL: 'reduce_smin_i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
-; SIZE-NEXT: Cost Model: Found an estimated cost of 6 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V2 = call i64 @llvm.vector.reduce.smin.v2i64(<2 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V4 = call i64 @llvm.vector.reduce.smin.v4i64(<4 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V8 = call i64 @llvm.vector.reduce.smin.v8i64(<8 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V16 = call i64 @llvm.vector.reduce.smin.v16i64(<16 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %V32 = call i64 @llvm.vector.reduce.smin.v32i64(<32 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %V64 = call i64 @llvm.vector.reduce.smin.v64i64(<64 x i64> undef)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 5 for instruction: %V128 = call i64 @llvm.vector.reduce.smin.v128i64(<128 x i64> undef)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 undef
;
%V1 = call i64 @llvm.vector.reduce.smin.v1i64(<1 x i64> undef)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
index e42dc889f1ba0..0dd3e3cf53e1c 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-fp.ll
@@ -456,7 +456,7 @@ define half @vreduce_fmin_nxv1f16(<vscale x 1 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -469,7 +469,7 @@ define half @vreduce_fmin_nxv1f16_nonans(<vscale x 1 x half> %v) #0 {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f16_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call nnan half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -482,7 +482,7 @@ define half @vreduce_fmin_nxv1f16_nonans_noinfs(<vscale x 1 x half> %v) #1 {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f16_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call nnan ninf half @llvm.vector.reduce.fmin.nxv1f16(<vscale x 1 x half> %v)
@@ -497,7 +497,7 @@ define half @vreduce_fmin_nxv2f16(<vscale x 2 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv2f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmin.nxv2f16(<vscale x 2 x half> %v)
@@ -512,7 +512,7 @@ define half @vreduce_fmin_nxv4f16(<vscale x 4 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv4f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmin.nxv4f16(<vscale x 4 x half> %v)
@@ -523,11 +523,11 @@ declare half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half>)
define half @vreduce_fmin_nxv64f16(<vscale x 64 x half> %v) {
; CHECK-LABEL: 'vreduce_fmin_nxv64f16'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv64f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmin.nxv64f16(<vscale x 64 x half> %v)
@@ -542,7 +542,7 @@ define float @vreduce_fmin_nxv1f32(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -555,7 +555,7 @@ define float @vreduce_fmin_nxv1f32_nonans(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f32_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call nnan float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -568,7 +568,7 @@ define float @vreduce_fmin_nxv1f32_nonans_noinfs(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f32_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call nnan ninf float @llvm.vector.reduce.fmin.nxv1f32(<vscale x 1 x float> %v)
@@ -583,7 +583,7 @@ define float @vreduce_fmin_nxv2f32(<vscale x 2 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv2f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmin.nxv2f32(<vscale x 2 x float> %v)
@@ -598,7 +598,7 @@ define float @vreduce_fmin_nxv4f32(<vscale x 4 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv4f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmin.nxv4f32(<vscale x 4 x float> %v)
@@ -609,11 +609,11 @@ declare float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float>)
define float @vreduce_fmin_nxv32f32(<vscale x 32 x float> %v) {
; CHECK-LABEL: 'vreduce_fmin_nxv32f32'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv32f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmin.nxv32f32(<vscale x 32 x float> %v)
@@ -628,7 +628,7 @@ define double @vreduce_fmin_nxv1f64(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -641,7 +641,7 @@ define double @vreduce_fmin_nxv1f64_nonans(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f64_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call nnan double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -654,7 +654,7 @@ define double @vreduce_fmin_nxv1f64_nonans_noinfs(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv1f64_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call nnan ninf double @llvm.vector.reduce.fmin.nxv1f64(<vscale x 1 x double> %v)
@@ -669,7 +669,7 @@ define double @vreduce_fmin_nxv2f64(<vscale x 2 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv2f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmin.nxv2f64(<vscale x 2 x double> %v)
@@ -684,7 +684,7 @@ define double @vreduce_fmin_nxv4f64(<vscale x 4 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv4f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmin.nxv4f64(<vscale x 4 x double> %v)
@@ -695,11 +695,11 @@ declare double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double>)
define double @vreduce_fmin_nxv16f64(<vscale x 16 x double> %v) {
; CHECK-LABEL: 'vreduce_fmin_nxv16f64'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmin_nxv16f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmin.nxv16f64(<vscale x 16 x double> %v)
@@ -714,7 +714,7 @@ define half @vreduce_fmax_nxv1f16(<vscale x 1 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -727,7 +727,7 @@ define half @vreduce_fmax_nxv1f16_nonans(<vscale x 1 x half> %v) #0 {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f16_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call nnan half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -740,7 +740,7 @@ define half @vreduce_fmax_nxv1f16_nonans_noinfs(<vscale x 1 x half> %v) #1 {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f16_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call nnan ninf half @llvm.vector.reduce.fmax.nxv1f16(<vscale x 1 x half> %v)
@@ -755,7 +755,7 @@ define half @vreduce_fmax_nxv2f16(<vscale x 2 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv2f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmax.nxv2f16(<vscale x 2 x half> %v)
@@ -770,7 +770,7 @@ define half @vreduce_fmax_nxv4f16(<vscale x 4 x half> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv4f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmax.nxv4f16(<vscale x 4 x half> %v)
@@ -781,11 +781,11 @@ declare half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half>)
define half @vreduce_fmax_nxv64f16(<vscale x 64 x half> %v) {
; CHECK-LABEL: 'vreduce_fmax_nxv64f16'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 16 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 10 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret half %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv64f16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret half %red
;
%red = call half @llvm.vector.reduce.fmax.nxv64f16(<vscale x 64 x half> %v)
@@ -800,7 +800,7 @@ define float @vreduce_fmax_nxv1f32(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -813,7 +813,7 @@ define float @vreduce_fmax_nxv1f32_nonans(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f32_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call nnan float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -826,7 +826,7 @@ define float @vreduce_fmax_nxv1f32_nonans_noinfs(<vscale x 1 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f32_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call nnan ninf float @llvm.vector.reduce.fmax.nxv1f32(<vscale x 1 x float> %v)
@@ -841,7 +841,7 @@ define float @vreduce_fmax_nxv2f32(<vscale x 2 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv2f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmax.nxv2f32(<vscale x 2 x float> %v)
@@ -856,7 +856,7 @@ define float @vreduce_fmax_nxv4f32(<vscale x 4 x float> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv4f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmax.nxv4f32(<vscale x 4 x float> %v)
@@ -867,11 +867,11 @@ declare float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float>)
define float @vreduce_fmax_nxv32f32(<vscale x 32 x float> %v) {
; CHECK-LABEL: 'vreduce_fmax_nxv32f32'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 15 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 9 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret float %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv32f32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret float %red
;
%red = call float @llvm.vector.reduce.fmax.nxv32f32(<vscale x 32 x float> %v)
@@ -886,7 +886,7 @@ define double @vreduce_fmax_nxv1f64(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -899,7 +899,7 @@ define double @vreduce_fmax_nxv1f64_nonans(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f64_nonans'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call nnan double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -912,7 +912,7 @@ define double @vreduce_fmax_nxv1f64_nonans_noinfs(<vscale x 1 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv1f64_nonans_noinfs'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call nnan ninf double @llvm.vector.reduce.fmax.nxv1f64(<vscale x 1 x double> %v)
@@ -927,7 +927,7 @@ define double @vreduce_fmax_nxv2f64(<vscale x 2 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv2f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmax.nxv2f64(<vscale x 2 x double> %v)
@@ -942,7 +942,7 @@ define double @vreduce_fmax_nxv4f64(<vscale x 4 x double> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv4f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmax.nxv4f64(<vscale x 4 x double> %v)
@@ -953,11 +953,11 @@ declare double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double>)
define double @vreduce_fmax_nxv16f64(<vscale x 16 x double> %v) {
; CHECK-LABEL: 'vreduce_fmax_nxv16f64'
-; CHECK-NEXT: Cost Model: Found an estimated cost of 14 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
+; CHECK-NEXT: Cost Model: Found an estimated cost of 8 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret double %red
;
; SIZE-LABEL: 'vreduce_fmax_nxv16f64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 4 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret double %red
;
%red = call double @llvm.vector.reduce.fmax.nxv16f64(<vscale x 16 x double> %v)
diff --git a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
index 2ffaf2dd22dce..80efe912c8698 100644
--- a/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
+++ b/llvm/test/Analysis/CostModel/RISCV/reduce-scalable-int.ll
@@ -27,7 +27,7 @@ define signext i8 @vreduce_umax_nxv1i8(<vscale x 1 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv1i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umax.nxv1i8(<vscale x 1 x i8> %v)
@@ -42,7 +42,7 @@ define signext i8 @vreduce_smax_nxv1i8(<vscale x 1 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv1i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smax.nxv1i8(<vscale x 1 x i8> %v)
@@ -57,7 +57,7 @@ define signext i8 @vreduce_umin_nxv1i8(<vscale x 1 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv1i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umin.nxv1i8(<vscale x 1 x i8> %v)
@@ -72,7 +72,7 @@ define signext i8 @vreduce_smin_nxv1i8(<vscale x 1 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv1i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smin.nxv1i8(<vscale x 1 x i8> %v)
@@ -147,7 +147,7 @@ define signext i8 @vreduce_umax_nxv2i8(<vscale x 2 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv2i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umax.nxv2i8(<vscale x 2 x i8> %v)
@@ -162,7 +162,7 @@ define signext i8 @vreduce_smax_nxv2i8(<vscale x 2 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv2i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smax.nxv2i8(<vscale x 2 x i8> %v)
@@ -177,7 +177,7 @@ define signext i8 @vreduce_umin_nxv2i8(<vscale x 2 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv2i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umin.nxv2i8(<vscale x 2 x i8> %v)
@@ -192,7 +192,7 @@ define signext i8 @vreduce_smin_nxv2i8(<vscale x 2 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv2i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smin.nxv2i8(<vscale x 2 x i8> %v)
@@ -267,7 +267,7 @@ define signext i8 @vreduce_umax_nxv4i8(<vscale x 4 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv4i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umax.nxv4i8(<vscale x 4 x i8> %v)
@@ -282,7 +282,7 @@ define signext i8 @vreduce_smax_nxv4i8(<vscale x 4 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv4i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smax.nxv4i8(<vscale x 4 x i8> %v)
@@ -297,7 +297,7 @@ define signext i8 @vreduce_umin_nxv4i8(<vscale x 4 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv4i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.umin.nxv4i8(<vscale x 4 x i8> %v)
@@ -312,7 +312,7 @@ define signext i8 @vreduce_smin_nxv4i8(<vscale x 4 x i8> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i8 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv4i8'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i8 %red
;
%red = call i8 @llvm.vector.reduce.smin.nxv4i8(<vscale x 4 x i8> %v)
@@ -419,7 +419,7 @@ define signext i16 @vreduce_umax_nxv1i16(<vscale x 1 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv1i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umax.nxv1i16(<vscale x 1 x i16> %v)
@@ -434,7 +434,7 @@ define signext i16 @vreduce_smax_nxv1i16(<vscale x 1 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv1i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smax.nxv1i16(<vscale x 1 x i16> %v)
@@ -449,7 +449,7 @@ define signext i16 @vreduce_umin_nxv1i16(<vscale x 1 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv1i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umin.nxv1i16(<vscale x 1 x i16> %v)
@@ -464,7 +464,7 @@ define signext i16 @vreduce_smin_nxv1i16(<vscale x 1 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv1i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smin.nxv1i16(<vscale x 1 x i16> %v)
@@ -571,7 +571,7 @@ define signext i16 @vreduce_umax_nxv2i16(<vscale x 2 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv2i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umax.nxv2i16(<vscale x 2 x i16> %v)
@@ -586,7 +586,7 @@ define signext i16 @vreduce_smax_nxv2i16(<vscale x 2 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv2i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smax.nxv2i16(<vscale x 2 x i16> %v)
@@ -601,7 +601,7 @@ define signext i16 @vreduce_umin_nxv2i16(<vscale x 2 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv2i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umin.nxv2i16(<vscale x 2 x i16> %v)
@@ -616,7 +616,7 @@ define signext i16 @vreduce_smin_nxv2i16(<vscale x 2 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv2i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smin.nxv2i16(<vscale x 2 x i16> %v)
@@ -723,7 +723,7 @@ define signext i16 @vreduce_umax_nxv4i16(<vscale x 4 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv4i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umax.nxv4i16(<vscale x 4 x i16> %v)
@@ -738,7 +738,7 @@ define signext i16 @vreduce_smax_nxv4i16(<vscale x 4 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv4i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smax.nxv4i16(<vscale x 4 x i16> %v)
@@ -753,7 +753,7 @@ define signext i16 @vreduce_umin_nxv4i16(<vscale x 4 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv4i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.umin.nxv4i16(<vscale x 4 x i16> %v)
@@ -768,7 +768,7 @@ define signext i16 @vreduce_smin_nxv4i16(<vscale x 4 x i16> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i16 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv4i16'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i16 %red
;
%red = call i16 @llvm.vector.reduce.smin.nxv4i16(<vscale x 4 x i16> %v)
@@ -875,7 +875,7 @@ define signext i32 @vreduce_umax_nxv1i32(<vscale x 1 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv1i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umax.nxv1i32(<vscale x 1 x i32> %v)
@@ -890,7 +890,7 @@ define signext i32 @vreduce_smax_nxv1i32(<vscale x 1 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv1i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smax.nxv1i32(<vscale x 1 x i32> %v)
@@ -905,7 +905,7 @@ define signext i32 @vreduce_umin_nxv1i32(<vscale x 1 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv1i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umin.nxv1i32(<vscale x 1 x i32> %v)
@@ -920,7 +920,7 @@ define signext i32 @vreduce_smin_nxv1i32(<vscale x 1 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv1i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smin.nxv1i32(<vscale x 1 x i32> %v)
@@ -1027,7 +1027,7 @@ define signext i32 @vreduce_umax_nxv2i32(<vscale x 2 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv2i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umax.nxv2i32(<vscale x 2 x i32> %v)
@@ -1042,7 +1042,7 @@ define signext i32 @vreduce_smax_nxv2i32(<vscale x 2 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv2i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smax.nxv2i32(<vscale x 2 x i32> %v)
@@ -1057,7 +1057,7 @@ define signext i32 @vreduce_umin_nxv2i32(<vscale x 2 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv2i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umin.nxv2i32(<vscale x 2 x i32> %v)
@@ -1072,7 +1072,7 @@ define signext i32 @vreduce_smin_nxv2i32(<vscale x 2 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv2i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smin.nxv2i32(<vscale x 2 x i32> %v)
@@ -1179,7 +1179,7 @@ define signext i32 @vreduce_umax_nxv4i32(<vscale x 4 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv4i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umax.nxv4i32(<vscale x 4 x i32> %v)
@@ -1194,7 +1194,7 @@ define signext i32 @vreduce_smax_nxv4i32(<vscale x 4 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv4i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smax.nxv4i32(<vscale x 4 x i32> %v)
@@ -1209,7 +1209,7 @@ define signext i32 @vreduce_umin_nxv4i32(<vscale x 4 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv4i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.umin.nxv4i32(<vscale x 4 x i32> %v)
@@ -1224,7 +1224,7 @@ define signext i32 @vreduce_smin_nxv4i32(<vscale x 4 x i32> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i32 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv4i32'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i32 %red
;
%red = call i32 @llvm.vector.reduce.smin.nxv4i32(<vscale x 4 x i32> %v)
@@ -1331,7 +1331,7 @@ define i64 @vreduce_umax_nxv1i64(<vscale x 1 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv1i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umax.nxv1i64(<vscale x 1 x i64> %v)
@@ -1346,7 +1346,7 @@ define i64 @vreduce_smax_nxv1i64(<vscale x 1 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv1i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smax.nxv1i64(<vscale x 1 x i64> %v)
@@ -1361,7 +1361,7 @@ define i64 @vreduce_umin_nxv1i64(<vscale x 1 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv1i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umin.nxv1i64(<vscale x 1 x i64> %v)
@@ -1376,7 +1376,7 @@ define i64 @vreduce_smin_nxv1i64(<vscale x 1 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv1i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smin.nxv1i64(<vscale x 1 x i64> %v)
@@ -1483,7 +1483,7 @@ define i64 @vreduce_umax_nxv2i64(<vscale x 2 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv2i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umax.nxv2i64(<vscale x 2 x i64> %v)
@@ -1498,7 +1498,7 @@ define i64 @vreduce_smax_nxv2i64(<vscale x 2 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv2i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smax.nxv2i64(<vscale x 2 x i64> %v)
@@ -1513,7 +1513,7 @@ define i64 @vreduce_umin_nxv2i64(<vscale x 2 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv2i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umin.nxv2i64(<vscale x 2 x i64> %v)
@@ -1528,7 +1528,7 @@ define i64 @vreduce_smin_nxv2i64(<vscale x 2 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv2i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smin.nxv2i64(<vscale x 2 x i64> %v)
@@ -1635,7 +1635,7 @@ define i64 @vreduce_umax_nxv4i64(<vscale x 4 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umax_nxv4i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umax.nxv4i64(<vscale x 4 x i64> %v)
@@ -1650,7 +1650,7 @@ define i64 @vreduce_smax_nxv4i64(<vscale x 4 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smax_nxv4i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smax.nxv4i64(<vscale x 4 x i64> %v)
@@ -1665,7 +1665,7 @@ define i64 @vreduce_umin_nxv4i64(<vscale x 4 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_umin_nxv4i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.umin.nxv4i64(<vscale x 4 x i64> %v)
@@ -1680,7 +1680,7 @@ define i64 @vreduce_smin_nxv4i64(<vscale x 4 x i64> %v) {
; CHECK-NEXT: Cost Model: Found an estimated cost of 0 for instruction: ret i64 %red
;
; SIZE-LABEL: 'vreduce_smin_nxv4i64'
-; SIZE-NEXT: Cost Model: Found an estimated cost of 3 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
+; SIZE-NEXT: Cost Model: Found an estimated cost of 2 for instruction: %red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
; SIZE-NEXT: Cost Model: Found an estimated cost of 1 for instruction: ret i64 %red
;
%red = call i64 @llvm.vector.reduce.smin.nxv4i64(<vscale x 4 x i64> %v)
>From edbd93d3706b806e7c91dfd2c359476488dfcc96 Mon Sep 17 00:00:00 2001
From: Nick Desaulniers <nickdesaulniers at users.noreply.github.com>
Date: Thu, 1 Feb 2024 13:09:33 -0800
Subject: [PATCH 08/12] [libc][stdbit] fix return types (#80337)
All of the functions I've previously implemented return an unsigned int; not
the parameter type.
---
libc/include/llvm-libc-macros/stdbit-macros.h | 16 +++----
libc/spec/stdc.td | 16 +++----
libc/src/stdbit/stdc_leading_ones_uc.cpp | 4 +-
libc/src/stdbit/stdc_leading_ones_uc.h | 2 +-
libc/src/stdbit/stdc_leading_ones_ul.cpp | 4 +-
libc/src/stdbit/stdc_leading_ones_ul.h | 2 +-
libc/src/stdbit/stdc_leading_ones_ull.cpp | 4 +-
libc/src/stdbit/stdc_leading_ones_ull.h | 2 +-
libc/src/stdbit/stdc_leading_ones_us.cpp | 5 +-
libc/src/stdbit/stdc_leading_ones_us.h | 2 +-
libc/src/stdbit/stdc_leading_zeros_uc.cpp | 5 +-
libc/src/stdbit/stdc_leading_zeros_uc.h | 2 +-
libc/src/stdbit/stdc_leading_zeros_ul.cpp | 5 +-
libc/src/stdbit/stdc_leading_zeros_ul.h | 2 +-
libc/src/stdbit/stdc_leading_zeros_ull.cpp | 4 +-
libc/src/stdbit/stdc_leading_zeros_ull.h | 2 +-
libc/src/stdbit/stdc_leading_zeros_us.cpp | 5 +-
libc/src/stdbit/stdc_leading_zeros_us.h | 2 +-
libc/test/include/stdbit_test.cpp | 48 ++++++++-----------
.../src/stdbit/stdc_leading_ones_uc_test.cpp | 4 +-
.../src/stdbit/stdc_leading_ones_ui_test.cpp | 2 +-
.../src/stdbit/stdc_leading_ones_ul_test.cpp | 4 +-
.../src/stdbit/stdc_leading_ones_ull_test.cpp | 4 +-
.../src/stdbit/stdc_leading_ones_us_test.cpp | 4 +-
.../src/stdbit/stdc_leading_zeros_uc_test.cpp | 4 +-
.../src/stdbit/stdc_leading_zeros_ul_test.cpp | 4 +-
.../stdbit/stdc_leading_zeros_ull_test.cpp | 4 +-
.../src/stdbit/stdc_leading_zeros_us_test.cpp | 4 +-
28 files changed, 77 insertions(+), 89 deletions(-)
diff --git a/libc/include/llvm-libc-macros/stdbit-macros.h b/libc/include/llvm-libc-macros/stdbit-macros.h
index 5355be2832fbb..cc964a5268b0d 100644
--- a/libc/include/llvm-libc-macros/stdbit-macros.h
+++ b/libc/include/llvm-libc-macros/stdbit-macros.h
@@ -10,34 +10,34 @@
#define __LLVM_LIBC_MACROS_STDBIT_MACROS_H
#ifdef __cplusplus
-inline unsigned char stdc_leading_zeros(unsigned char x) {
+inline unsigned stdc_leading_zeros(unsigned char x) {
return stdc_leading_zeros_uc(x);
}
-inline unsigned short stdc_leading_zeros(unsigned short x) {
+inline unsigned stdc_leading_zeros(unsigned short x) {
return stdc_leading_zeros_us(x);
}
inline unsigned stdc_leading_zeros(unsigned x) {
return stdc_leading_zeros_ui(x);
}
-inline unsigned long stdc_leading_zeros(unsigned long x) {
+inline unsigned stdc_leading_zeros(unsigned long x) {
return stdc_leading_zeros_ul(x);
}
-inline unsigned long long stdc_leading_zeros(unsigned long long x) {
+inline unsigned stdc_leading_zeros(unsigned long long x) {
return stdc_leading_zeros_ull(x);
}
-inline unsigned char stdc_leading_ones(unsigned char x) {
+inline unsigned stdc_leading_ones(unsigned char x) {
return stdc_leading_ones_uc(x);
}
-inline unsigned short stdc_leading_ones(unsigned short x) {
+inline unsigned stdc_leading_ones(unsigned short x) {
return stdc_leading_ones_us(x);
}
inline unsigned stdc_leading_ones(unsigned x) {
return stdc_leading_ones_ui(x);
}
-inline unsigned long stdc_leading_ones(unsigned long x) {
+inline unsigned stdc_leading_ones(unsigned long x) {
return stdc_leading_ones_ul(x);
}
-inline unsigned long long stdc_leading_ones(unsigned long long x) {
+inline unsigned stdc_leading_ones(unsigned long long x) {
return stdc_leading_ones_ull(x);
}
#else
diff --git a/libc/spec/stdc.td b/libc/spec/stdc.td
index cb23a6700d913..6ff2c7c613696 100644
--- a/libc/spec/stdc.td
+++ b/libc/spec/stdc.td
@@ -775,16 +775,16 @@ def StdC : StandardSpec<"stdc"> {
[], // Types
[], // Enumerations
[
- FunctionSpec<"stdc_leading_zeros_uc", RetValSpec<UnsignedCharType>, [ArgSpec<UnsignedCharType>]>,
- FunctionSpec<"stdc_leading_zeros_us", RetValSpec<UnsignedShortType>, [ArgSpec<UnsignedShortType>]>,
+ FunctionSpec<"stdc_leading_zeros_uc", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedCharType>]>,
+ FunctionSpec<"stdc_leading_zeros_us", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedShortType>]>,
FunctionSpec<"stdc_leading_zeros_ui", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedIntType>]>,
- FunctionSpec<"stdc_leading_zeros_ul", RetValSpec<UnsignedLongType>, [ArgSpec<UnsignedLongType>]>,
- FunctionSpec<"stdc_leading_zeros_ull", RetValSpec<UnsignedLongLongType>, [ArgSpec<UnsignedLongLongType>]>,
- FunctionSpec<"stdc_leading_ones_uc", RetValSpec<UnsignedCharType>, [ArgSpec<UnsignedCharType>]>,
- FunctionSpec<"stdc_leading_ones_us", RetValSpec<UnsignedShortType>, [ArgSpec<UnsignedShortType>]>,
+ FunctionSpec<"stdc_leading_zeros_ul", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedLongType>]>,
+ FunctionSpec<"stdc_leading_zeros_ull", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedLongLongType>]>,
+ FunctionSpec<"stdc_leading_ones_uc", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedCharType>]>,
+ FunctionSpec<"stdc_leading_ones_us", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedShortType>]>,
FunctionSpec<"stdc_leading_ones_ui", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedIntType>]>,
- FunctionSpec<"stdc_leading_ones_ul", RetValSpec<UnsignedLongType>, [ArgSpec<UnsignedLongType>]>,
- FunctionSpec<"stdc_leading_ones_ull", RetValSpec<UnsignedLongLongType>, [ArgSpec<UnsignedLongLongType>]>
+ FunctionSpec<"stdc_leading_ones_ul", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedLongType>]>,
+ FunctionSpec<"stdc_leading_ones_ull", RetValSpec<UnsignedIntType>, [ArgSpec<UnsignedLongLongType>]>
] // Functions
>;
diff --git a/libc/src/stdbit/stdc_leading_ones_uc.cpp b/libc/src/stdbit/stdc_leading_ones_uc.cpp
index beae134eeb91a..69322d957dbf2 100644
--- a/libc/src/stdbit/stdc_leading_ones_uc.cpp
+++ b/libc/src/stdbit/stdc_leading_ones_uc.cpp
@@ -13,8 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned char, stdc_leading_ones_uc, (unsigned char value)) {
- return static_cast<unsigned char>(cpp::countl_one(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_ones_uc, (unsigned char value)) {
+ return static_cast<unsigned>(cpp::countl_one(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_uc.h b/libc/src/stdbit/stdc_leading_ones_uc.h
index 8303f3f7bd26e..fc4d2bdf623a6 100644
--- a/libc/src/stdbit/stdc_leading_ones_uc.h
+++ b/libc/src/stdbit/stdc_leading_ones_uc.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned char stdc_leading_ones_uc(unsigned char value);
+unsigned stdc_leading_ones_uc(unsigned char value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_ul.cpp b/libc/src/stdbit/stdc_leading_ones_ul.cpp
index ccfd9b9594e0b..5b2f61d43b027 100644
--- a/libc/src/stdbit/stdc_leading_ones_ul.cpp
+++ b/libc/src/stdbit/stdc_leading_ones_ul.cpp
@@ -13,8 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned long, stdc_leading_ones_ul, (unsigned long value)) {
- return static_cast<unsigned long>(cpp::countl_one(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_ones_ul, (unsigned long value)) {
+ return static_cast<unsigned>(cpp::countl_one(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_ul.h b/libc/src/stdbit/stdc_leading_ones_ul.h
index 7643bd1ba81d9..3441edf861361 100644
--- a/libc/src/stdbit/stdc_leading_ones_ul.h
+++ b/libc/src/stdbit/stdc_leading_ones_ul.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned long stdc_leading_ones_ul(unsigned long value);
+unsigned stdc_leading_ones_ul(unsigned long value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_ull.cpp b/libc/src/stdbit/stdc_leading_ones_ull.cpp
index 8cb1ff08c7b00..05660b9fed653 100644
--- a/libc/src/stdbit/stdc_leading_ones_ull.cpp
+++ b/libc/src/stdbit/stdc_leading_ones_ull.cpp
@@ -13,9 +13,9 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned long long, stdc_leading_ones_ull,
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_ones_ull,
(unsigned long long value)) {
- return static_cast<unsigned long long>(cpp::countl_one(value));
+ return static_cast<unsigned>(cpp::countl_one(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_ull.h b/libc/src/stdbit/stdc_leading_ones_ull.h
index 8ae8145109671..167eb732794d7 100644
--- a/libc/src/stdbit/stdc_leading_ones_ull.h
+++ b/libc/src/stdbit/stdc_leading_ones_ull.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned long long stdc_leading_ones_ull(unsigned long long value);
+unsigned stdc_leading_ones_ull(unsigned long long value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_us.cpp b/libc/src/stdbit/stdc_leading_ones_us.cpp
index 4b58cf5cee433..d93327db739b5 100644
--- a/libc/src/stdbit/stdc_leading_ones_us.cpp
+++ b/libc/src/stdbit/stdc_leading_ones_us.cpp
@@ -13,9 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned short, stdc_leading_ones_us,
- (unsigned short value)) {
- return static_cast<unsigned short>(cpp::countl_one(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_ones_us, (unsigned short value)) {
+ return static_cast<unsigned>(cpp::countl_one(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_ones_us.h b/libc/src/stdbit/stdc_leading_ones_us.h
index 583a479f79a81..2a552d784b195 100644
--- a/libc/src/stdbit/stdc_leading_ones_us.h
+++ b/libc/src/stdbit/stdc_leading_ones_us.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned short stdc_leading_ones_us(unsigned short value);
+unsigned stdc_leading_ones_us(unsigned short value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_uc.cpp b/libc/src/stdbit/stdc_leading_zeros_uc.cpp
index 8c27043e055bf..dbd2d7662c44f 100644
--- a/libc/src/stdbit/stdc_leading_zeros_uc.cpp
+++ b/libc/src/stdbit/stdc_leading_zeros_uc.cpp
@@ -13,9 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned char, stdc_leading_zeros_uc,
- (unsigned char value)) {
- return static_cast<unsigned char>(cpp::countl_zero(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_zeros_uc, (unsigned char value)) {
+ return static_cast<unsigned>(cpp::countl_zero(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_uc.h b/libc/src/stdbit/stdc_leading_zeros_uc.h
index 0622e72b0683e..1edca66e13448 100644
--- a/libc/src/stdbit/stdc_leading_zeros_uc.h
+++ b/libc/src/stdbit/stdc_leading_zeros_uc.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned char stdc_leading_zeros_uc(unsigned char value);
+unsigned stdc_leading_zeros_uc(unsigned char value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_ul.cpp b/libc/src/stdbit/stdc_leading_zeros_ul.cpp
index cf21ff449a8bf..0d5ff886fe991 100644
--- a/libc/src/stdbit/stdc_leading_zeros_ul.cpp
+++ b/libc/src/stdbit/stdc_leading_zeros_ul.cpp
@@ -13,9 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned long, stdc_leading_zeros_ul,
- (unsigned long value)) {
- return static_cast<unsigned long>(cpp::countl_zero(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_zeros_ul, (unsigned long value)) {
+ return static_cast<unsigned>(cpp::countl_zero(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_ul.h b/libc/src/stdbit/stdc_leading_zeros_ul.h
index a1f36299931dc..5bac6df1b14dc 100644
--- a/libc/src/stdbit/stdc_leading_zeros_ul.h
+++ b/libc/src/stdbit/stdc_leading_zeros_ul.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned long stdc_leading_zeros_ul(unsigned long value);
+unsigned stdc_leading_zeros_ul(unsigned long value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_ull.cpp b/libc/src/stdbit/stdc_leading_zeros_ull.cpp
index d488559bbd14c..b23cea7c3ee44 100644
--- a/libc/src/stdbit/stdc_leading_zeros_ull.cpp
+++ b/libc/src/stdbit/stdc_leading_zeros_ull.cpp
@@ -13,9 +13,9 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned long long, stdc_leading_zeros_ull,
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_zeros_ull,
(unsigned long long value)) {
- return static_cast<unsigned long long>(cpp::countl_zero(value));
+ return static_cast<unsigned>(cpp::countl_zero(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_ull.h b/libc/src/stdbit/stdc_leading_zeros_ull.h
index b05855b296af0..e13624345fc78 100644
--- a/libc/src/stdbit/stdc_leading_zeros_ull.h
+++ b/libc/src/stdbit/stdc_leading_zeros_ull.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned long long stdc_leading_zeros_ull(unsigned long long value);
+unsigned stdc_leading_zeros_ull(unsigned long long value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_us.cpp b/libc/src/stdbit/stdc_leading_zeros_us.cpp
index 84df2b842d695..424fdc7282563 100644
--- a/libc/src/stdbit/stdc_leading_zeros_us.cpp
+++ b/libc/src/stdbit/stdc_leading_zeros_us.cpp
@@ -13,9 +13,8 @@
namespace LIBC_NAMESPACE {
-LLVM_LIBC_FUNCTION(unsigned short, stdc_leading_zeros_us,
- (unsigned short value)) {
- return static_cast<unsigned short>(cpp::countl_zero(value));
+LLVM_LIBC_FUNCTION(unsigned, stdc_leading_zeros_us, (unsigned short value)) {
+ return static_cast<unsigned>(cpp::countl_zero(value));
}
} // namespace LIBC_NAMESPACE
diff --git a/libc/src/stdbit/stdc_leading_zeros_us.h b/libc/src/stdbit/stdc_leading_zeros_us.h
index c0f62e2be2f02..6a3969002daf5 100644
--- a/libc/src/stdbit/stdc_leading_zeros_us.h
+++ b/libc/src/stdbit/stdc_leading_zeros_us.h
@@ -11,7 +11,7 @@
namespace LIBC_NAMESPACE {
-unsigned short stdc_leading_zeros_us(unsigned short value);
+unsigned stdc_leading_zeros_us(unsigned short value);
} // namespace LIBC_NAMESPACE
diff --git a/libc/test/include/stdbit_test.cpp b/libc/test/include/stdbit_test.cpp
index 5e45f971cb22b..6e48b0f2f7594 100644
--- a/libc/test/include/stdbit_test.cpp
+++ b/libc/test/include/stdbit_test.cpp
@@ -23,40 +23,32 @@
* enabled.
*/
extern "C" {
-unsigned char stdc_leading_zeros_uc(unsigned char) noexcept { return 0xAA; }
-unsigned short stdc_leading_zeros_us(unsigned short) noexcept { return 0xAB; }
-unsigned stdc_leading_zeros_ui(unsigned) noexcept { return 0xAC; }
-unsigned long stdc_leading_zeros_ul(unsigned long) noexcept { return 0xAD; }
-unsigned long long stdc_leading_zeros_ull(unsigned long long) noexcept {
- return 0xAF;
-}
-unsigned char stdc_leading_ones_uc(unsigned char) noexcept { return 0xBA; }
-unsigned short stdc_leading_ones_us(unsigned short) noexcept { return 0xBB; }
-unsigned stdc_leading_ones_ui(unsigned) noexcept { return 0xBC; }
-unsigned long stdc_leading_ones_ul(unsigned long) noexcept { return 0xBD; }
-unsigned long long stdc_leading_ones_ull(unsigned long long) noexcept {
- return 0xBF;
-}
+unsigned stdc_leading_zeros_uc(unsigned char) noexcept { return 0xAAU; }
+unsigned stdc_leading_zeros_us(unsigned short) noexcept { return 0xABU; }
+unsigned stdc_leading_zeros_ui(unsigned) noexcept { return 0xACU; }
+unsigned stdc_leading_zeros_ul(unsigned long) noexcept { return 0xADU; }
+unsigned stdc_leading_zeros_ull(unsigned long long) noexcept { return 0xAFU; }
+unsigned stdc_leading_ones_uc(unsigned char) noexcept { return 0xBAU; }
+unsigned stdc_leading_ones_us(unsigned short) noexcept { return 0xBBU; }
+unsigned stdc_leading_ones_ui(unsigned) noexcept { return 0xBCU; }
+unsigned stdc_leading_ones_ul(unsigned long) noexcept { return 0xBDU; }
+unsigned stdc_leading_ones_ull(unsigned long long) noexcept { return 0xBFU; }
}
#include "include/llvm-libc-macros/stdbit-macros.h"
TEST(LlvmLibcStdbitTest, TypeGenericMacroLeadingZeros) {
- EXPECT_EQ(stdc_leading_zeros(static_cast<unsigned char>(0U)),
- static_cast<unsigned char>(0xAA));
- EXPECT_EQ(stdc_leading_zeros(static_cast<unsigned short>(0U)),
- static_cast<unsigned short>(0xAB));
- EXPECT_EQ(stdc_leading_zeros(0U), static_cast<unsigned>(0xAC));
- EXPECT_EQ(stdc_leading_zeros(0UL), static_cast<unsigned long>(0xAD));
- EXPECT_EQ(stdc_leading_zeros(0ULL), static_cast<unsigned long long>(0xAF));
+ EXPECT_EQ(stdc_leading_zeros(static_cast<unsigned char>(0U)), 0xAAU);
+ EXPECT_EQ(stdc_leading_zeros(static_cast<unsigned short>(0U)), 0xABU);
+ EXPECT_EQ(stdc_leading_zeros(0U), 0xACU);
+ EXPECT_EQ(stdc_leading_zeros(0UL), 0xADU);
+ EXPECT_EQ(stdc_leading_zeros(0ULL), 0xAFU);
}
TEST(LlvmLibcStdbitTest, TypeGenericMacroLeadingOnes) {
- EXPECT_EQ(stdc_leading_ones(static_cast<unsigned char>(0U)),
- static_cast<unsigned char>(0xBA));
- EXPECT_EQ(stdc_leading_ones(static_cast<unsigned short>(0U)),
- static_cast<unsigned short>(0xBB));
- EXPECT_EQ(stdc_leading_ones(0U), static_cast<unsigned>(0xBC));
- EXPECT_EQ(stdc_leading_ones(0UL), static_cast<unsigned long>(0xBD));
- EXPECT_EQ(stdc_leading_ones(0ULL), static_cast<unsigned long long>(0xBF));
+ EXPECT_EQ(stdc_leading_ones(static_cast<unsigned char>(0U)), 0xBAU);
+ EXPECT_EQ(stdc_leading_ones(static_cast<unsigned short>(0U)), 0xBBU);
+ EXPECT_EQ(stdc_leading_ones(0U), 0xBCU);
+ EXPECT_EQ(stdc_leading_ones(0UL), 0xBDU);
+ EXPECT_EQ(stdc_leading_ones(0ULL), 0xBFU);
}
diff --git a/libc/test/src/stdbit/stdc_leading_ones_uc_test.cpp b/libc/test/src/stdbit/stdc_leading_ones_uc_test.cpp
index 2883065df3771..5d32d92e327a3 100644
--- a/libc/test/src/stdbit/stdc_leading_ones_uc_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_ones_uc_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingOnesUcTest, All) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_uc(UCHAR_MAX),
- static_cast<unsigned char>(UCHAR_WIDTH));
+ static_cast<unsigned>(UCHAR_WIDTH));
}
TEST(LlvmLibcStdcLeadingOnesUcTest, ZeroHot) {
for (unsigned i = 0U; i != UCHAR_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_uc(~(1U << i)),
- static_cast<unsigned char>(UCHAR_WIDTH - i - 1));
+ UCHAR_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_ones_ui_test.cpp b/libc/test/src/stdbit/stdc_leading_ones_ui_test.cpp
index bae00573927ac..175a48b0d323e 100644
--- a/libc/test/src/stdbit/stdc_leading_ones_ui_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_ones_ui_test.cpp
@@ -18,5 +18,5 @@ TEST(LlvmLibcStdcLeadingOnesUiTest, All) {
TEST(LlvmLibcStdcLeadingOnesUiTest, ZeroHot) {
for (unsigned i = 0U; i != UINT_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_ui(~(1U << i)),
- static_cast<unsigned>(UINT_WIDTH - i - 1));
+ UINT_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_ones_ul_test.cpp b/libc/test/src/stdbit/stdc_leading_ones_ul_test.cpp
index bb7b9755ebcc5..b216f4c28acd2 100644
--- a/libc/test/src/stdbit/stdc_leading_ones_ul_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_ones_ul_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingOnesUlTest, All) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_ul(ULONG_MAX),
- static_cast<unsigned long>(ULONG_WIDTH));
+ static_cast<unsigned>(ULONG_WIDTH));
}
TEST(LlvmLibcStdcLeadingOnesUlTest, ZeroHot) {
for (unsigned i = 0U; i != ULONG_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_ul(~(1UL << i)),
- static_cast<unsigned long>(ULONG_WIDTH - i - 1));
+ ULONG_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_ones_ull_test.cpp b/libc/test/src/stdbit/stdc_leading_ones_ull_test.cpp
index a8497a62bdd17..176376665bbd4 100644
--- a/libc/test/src/stdbit/stdc_leading_ones_ull_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_ones_ull_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingOnesUllTest, All) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_ull(ULLONG_MAX),
- static_cast<unsigned long long>(ULLONG_WIDTH));
+ static_cast<unsigned>(ULLONG_WIDTH));
}
TEST(LlvmLibcStdcLeadingOnesUllTest, ZeroHot) {
for (unsigned i = 0U; i != ULLONG_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_ull(~(1ULL << i)),
- static_cast<unsigned long long>(ULLONG_WIDTH - i - 1));
+ ULLONG_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_ones_us_test.cpp b/libc/test/src/stdbit/stdc_leading_ones_us_test.cpp
index 40c34126b76dd..91a125370ec15 100644
--- a/libc/test/src/stdbit/stdc_leading_ones_us_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_ones_us_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingOnesUsTest, All) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_us(USHRT_MAX),
- static_cast<unsigned short>(USHRT_WIDTH));
+ static_cast<unsigned>(USHRT_WIDTH));
}
TEST(LlvmLibcStdcLeadingOnesUsTest, ZeroHot) {
for (unsigned i = 0U; i != USHRT_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_ones_us(~(1U << i)),
- static_cast<unsigned short>(USHRT_WIDTH - i - 1));
+ USHRT_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_zeros_uc_test.cpp b/libc/test/src/stdbit/stdc_leading_zeros_uc_test.cpp
index 4e2e1db3bd72b..3d555072927ac 100644
--- a/libc/test/src/stdbit/stdc_leading_zeros_uc_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_zeros_uc_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingZerosUcTest, Zero) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_uc(0U),
- static_cast<unsigned char>(UCHAR_WIDTH));
+ static_cast<unsigned>(UCHAR_WIDTH));
}
TEST(LlvmLibcStdcLeadingZerosUcTest, OneHot) {
for (unsigned i = 0U; i != UCHAR_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_uc(1U << i),
- static_cast<unsigned char>(UCHAR_WIDTH - i - 1));
+ UCHAR_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_zeros_ul_test.cpp b/libc/test/src/stdbit/stdc_leading_zeros_ul_test.cpp
index 9a73aece89acc..fd82cef548b6f 100644
--- a/libc/test/src/stdbit/stdc_leading_zeros_ul_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_zeros_ul_test.cpp
@@ -13,11 +13,11 @@
TEST(LlvmLibcStdcLeadingZerosUlTest, Zero) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_ul(0UL),
- static_cast<unsigned long>(ULONG_WIDTH));
+ static_cast<unsigned>(ULONG_WIDTH));
}
TEST(LlvmLibcStdcLeadingZerosUlTest, OneHot) {
for (unsigned i = 0U; i != ULONG_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_ul(1UL << i),
- static_cast<unsigned long>(ULONG_WIDTH - i - 1));
+ ULONG_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_zeros_ull_test.cpp b/libc/test/src/stdbit/stdc_leading_zeros_ull_test.cpp
index 9df2f015f9b6f..cdeeb30323004 100644
--- a/libc/test/src/stdbit/stdc_leading_zeros_ull_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_zeros_ull_test.cpp
@@ -13,11 +13,11 @@
TEST(LlvmLibcStdcLeadingZerosUllTest, Zero) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_ull(0ULL),
- static_cast<unsigned long long>(ULLONG_WIDTH));
+ static_cast<unsigned>(ULLONG_WIDTH));
}
TEST(LlvmLibcStdcLeadingZerosUllTest, OneHot) {
for (unsigned i = 0U; i != ULLONG_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_ull(1ULL << i),
- static_cast<unsigned long long>(ULLONG_WIDTH - i - 1));
+ ULLONG_WIDTH - i - 1U);
}
diff --git a/libc/test/src/stdbit/stdc_leading_zeros_us_test.cpp b/libc/test/src/stdbit/stdc_leading_zeros_us_test.cpp
index b396a740e43b7..afb418a24ad54 100644
--- a/libc/test/src/stdbit/stdc_leading_zeros_us_test.cpp
+++ b/libc/test/src/stdbit/stdc_leading_zeros_us_test.cpp
@@ -12,11 +12,11 @@
TEST(LlvmLibcStdcLeadingZerosUsTest, Zero) {
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_us(0U),
- static_cast<unsigned short>(USHRT_WIDTH));
+ static_cast<unsigned>(USHRT_WIDTH));
}
TEST(LlvmLibcStdcLeadingZerosUsTest, OneHot) {
for (unsigned i = 0U; i != USHRT_WIDTH; ++i)
EXPECT_EQ(LIBC_NAMESPACE::stdc_leading_zeros_us(1U << i),
- static_cast<unsigned short>(USHRT_WIDTH - i - 1));
+ USHRT_WIDTH - i - 1U);
}
>From a3a294bc3253f75498f88d9c8737f02c345712df Mon Sep 17 00:00:00 2001
From: Rahman Lavaee <rahmanl at google.com>
Date: Fri, 1 Dec 2023 09:27:35 +0000
Subject: [PATCH 09/12] [SHT_LLVM_BB_ADDR_MAP] Allow basic-block-sections and
labels be used together by decoupling the handling of the two features.
Today `-split-machine-functions` (MFS) and `-fbasic-block-sections={all,list}` cannot be combined with `-basic-block-sections=labels` (the labels option will be ignored).
The inconsistency comes from the way basic block address map -- the underlying mechanism for basic block labels -- encodes basic block addresses (https://lists.llvm.org/pipermail/llvm-dev/2020-July/143512.html). Specifically, basic block offsets are computed relative to the function begin symbol. This relies on functions being contiguous which is not the case for MFS and basic block section binaries. This means Propeller cannot use binary profiles collected from these binaries, which limits the applicability of Propeller for iterative optimization.
To make the BB address map feature work with BB section binaries, we propose modifying the encoding of the BB address map as follows.
The current encoding emits the address of each function and its number of basic blocks, followed by basic block entries for each basic block.
| Address of the function | //Address// | -> 8 bytes (pointer size)
| Number of basic blocks in this function |//NumBlocks// | -> ULEB128
| BB entry #1
| BB entry #2
| ...
| BB entry #//NumBlocks//
To make this work for basic block sections, we treat each basic block section similar to a function, except that basic block sections of the same function must be encapsulated in the same structure so we can map all of them to their single function.
We modify the encoding to first emit the number of basic block sections (BB ranges) in the function. Then we emit the address map of each basic block section section as before: the base address of the section, its number of blocks, and BB entries for its basic block. The first section in the BB address map is always the function entry section.
| Number of sections for this function | //NumBBRanges//) | -> ULEB128
| Section //0// begin address |//BaseAddress[0]// | -> 8 bytes (pointer size)
| Number of basic blocks in section //0// |//NumBlocks[0]// | -> ULEB128
| BB entries for Section //0//
.
.
.
| Section //K// begin address |//BaseAddress[K]// | -> 8 bytes (pointer size)
| Number of basic blocks in section //K// |//NumBlocks[K]// | -> ULEB128
| BB entries for Section //K//
The encoding of basic block entries remains as before with the minor change that each basic block offset is now computed relative to the begin symbol of its containing BB section.
This patch adds a new boolean codegen option `-basic-block-address-map`. Correspondingly, the front-end flag `-fbasic-block-address-map` and LLD flag `--lto-basic-block-address-map` are introduced.
Analogously, we add a new TargetOption field `BBAddrMap`. This means BB address maps are either generated for all functions in the compiling unit, or for none (depending on `TargetOptions::BBAddrMap`).
This patch disables the old `-fbasic-block-sections=labels` value option but does not remove it. A subsequent patch will remove the obsolete option.
We refactor the `BasicBlockSections` pass by separating the BB address map and BB sections handing to their own functions (named `handleBBAddrMap` and `handleBBSections`). `handleBBSections` renumbers basic blocks and places them in their assigned sections. `handleBBAddrMap` is invoked after `handleBBSections` (if requested) and only renumbers the blocks.
- New tests added:
# A codgen test `basic-block-labels-with-sections.ll` to exercise the combination of `-basic-block-address-map` with `-basic-block-sections=list`.
# A driver sanity test for the `-fbasic-block-address-map` option.
# An LLD test for testing the `--lto-basic-block-address-map` option. This reuses the LLVM IR from `lld/test/ELF/lto/basic-block-sections.ll`.
- Renamed and modified the two existing codegen tests for basic block address map (`basic-block-sections-labels-functions-sections.ll` and `basic-block-sections-labels.ll`)
Differential Revision: https://reviews.llvm.org/D97526
---
clang/include/clang/Basic/CodeGenOptions.def | 1 +
clang/include/clang/Driver/Options.td | 4 +
clang/lib/CodeGen/BackendUtil.cpp | 1 +
clang/lib/Driver/ToolChains/Clang.cpp | 11 +
clang/test/Driver/basic-block-address-map.c | 8 +
lld/ELF/Config.h | 1 +
lld/ELF/Driver.cpp | 3 +
lld/ELF/LTO.cpp | 2 +
lld/ELF/Options.td | 3 +
lld/test/ELF/lto/basic-block-address-map.ll | 28 ++
llvm/include/llvm/CodeGen/CommandFlags.h | 2 +
llvm/include/llvm/Object/ELFTypes.h | 138 ++++--
llvm/include/llvm/ObjectYAML/ELFYAML.h | 27 +-
llvm/include/llvm/Target/TargetOptions.h | 5 +-
llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp | 91 +++-
llvm/lib/CodeGen/BasicBlockSections.cpp | 39 +-
llvm/lib/CodeGen/CommandFlags.cpp | 7 +
llvm/lib/CodeGen/MachineFunction.cpp | 1 +
llvm/lib/CodeGen/TargetPassConfig.cpp | 28 +-
llvm/lib/Object/ELF.cpp | 126 +++--
llvm/lib/ObjectYAML/ELFEmitter.cpp | 58 ++-
llvm/lib/ObjectYAML/ELFYAML.cpp | 9 +-
...ic-block-address-map-function-sections.ll} | 1 +
...k-address-map-with-basic-block-sections.ll | 71 +++
.../X86/basic-block-address-map-with-mfs.ll | 89 ++++
...s-labels.ll => basic-block-address-map.ll} | 2 +
...ddrmap-disassemble-symbolize-operands.yaml | 120 ++---
.../elf-bbaddrmap-symbolize-relocatable.yaml | 22 +-
.../llvm-objdump/X86/elf-pgoanalysismap.yaml | 87 ++--
.../ELF/bb-addr-map-relocatable.test | 107 +++--
.../tools/llvm-readobj/ELF/bb-addr-map.test | 272 ++++-------
llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml | 216 +++------
llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml | 39 +-
llvm/tools/llvm-objdump/llvm-objdump.cpp | 194 +++++---
llvm/tools/llvm-readobj/ELFDumper.cpp | 17 +-
llvm/tools/obj2yaml/elf2yaml.cpp | 46 +-
llvm/unittests/Object/ELFObjectFileTest.cpp | 448 +++++++++++-------
llvm/unittests/Object/ELFTypesTest.cpp | 38 +-
38 files changed, 1420 insertions(+), 942 deletions(-)
create mode 100644 clang/test/Driver/basic-block-address-map.c
create mode 100644 lld/test/ELF/lto/basic-block-address-map.ll
rename llvm/test/CodeGen/X86/{basic-block-sections-labels-functions-sections.ll => basic-block-address-map-function-sections.ll} (94%)
create mode 100644 llvm/test/CodeGen/X86/basic-block-address-map-with-basic-block-sections.ll
create mode 100644 llvm/test/CodeGen/X86/basic-block-address-map-with-mfs.ll
rename llvm/test/CodeGen/X86/{basic-block-sections-labels.ll => basic-block-address-map.ll} (90%)
diff --git a/clang/include/clang/Basic/CodeGenOptions.def b/clang/include/clang/Basic/CodeGenOptions.def
index 7c0bfe3284961..6ad0503196256 100644
--- a/clang/include/clang/Basic/CodeGenOptions.def
+++ b/clang/include/clang/Basic/CodeGenOptions.def
@@ -96,6 +96,7 @@ ENUM_CODEGENOPT(InlineAsmDialect, InlineAsmDialectKind, 1, IAD_ATT)
CODEGENOPT(ForbidGuardVariables , 1, 0) ///< Issue errors if C++ guard variables
///< are required.
CODEGENOPT(FunctionSections , 1, 0) ///< Set when -ffunction-sections is enabled.
+CODEGENOPT(BBAddrMap , 1, 0) ///< Set when -fbasic-block-address-map is enabled.
CODEGENOPT(InstrumentFunctions , 1, 0) ///< Set when -finstrument-functions is
///< enabled.
CODEGENOPT(InstrumentFunctionsAfterInlining , 1, 0) ///< Set when
diff --git a/clang/include/clang/Driver/Options.td b/clang/include/clang/Driver/Options.td
index a4b35e370999e..73071a6648541 100644
--- a/clang/include/clang/Driver/Options.td
+++ b/clang/include/clang/Driver/Options.td
@@ -3988,6 +3988,10 @@ defm function_sections : BoolFOption<"function-sections",
PosFlag<SetTrue, [], [ClangOption, CC1Option],
"Place each function in its own section">,
NegFlag<SetFalse>>;
+defm basic_block_address_map : BoolFOption<"basic-block-address-map",
+ CodeGenOpts<"BBAddrMap">, DefaultFalse,
+ PosFlag<SetTrue, [], [CC1Option], "Emit the basic block address map section.">,
+ NegFlag<SetFalse>>;
def fbasic_block_sections_EQ : Joined<["-"], "fbasic-block-sections=">, Group<f_Group>,
Visibility<[ClangOption, CC1Option, CC1AsOption]>,
HelpText<"Place each function's basic blocks in unique sections (ELF Only)">,
diff --git a/clang/lib/CodeGen/BackendUtil.cpp b/clang/lib/CodeGen/BackendUtil.cpp
index 7877e20d77f77..e2511aa1b5c99 100644
--- a/clang/lib/CodeGen/BackendUtil.cpp
+++ b/clang/lib/CodeGen/BackendUtil.cpp
@@ -374,6 +374,7 @@ static bool initTargetOptions(DiagnosticsEngine &Diags,
LangOptions::FPModeKind::FPM_FastHonorPragmas);
Options.ApproxFuncFPMath = LangOpts.ApproxFunc;
+ Options.BBAddrMap = CodeGenOpts.BBAddrMap;
Options.BBSections =
llvm::StringSwitch<llvm::BasicBlockSection>(CodeGenOpts.BBSections)
.Case("all", llvm::BasicBlockSection::All)
diff --git a/clang/lib/Driver/ToolChains/Clang.cpp b/clang/lib/Driver/ToolChains/Clang.cpp
index 6a5f0b3b51b1f..7009248bc4f02 100644
--- a/clang/lib/Driver/ToolChains/Clang.cpp
+++ b/clang/lib/Driver/ToolChains/Clang.cpp
@@ -5944,6 +5944,17 @@ void Clang::ConstructJob(Compilation &C, const JobAction &JA,
CmdArgs.push_back("-ffunction-sections");
}
+ if (Arg *A = Args.getLastArg(options::OPT_fbasic_block_address_map,
+ options::OPT_fno_basic_block_address_map)) {
+ if (Triple.isX86() && Triple.isOSBinFormatELF()) {
+ if (A->getOption().matches(options::OPT_fbasic_block_address_map))
+ A->render(Args, CmdArgs);
+ } else {
+ D.Diag(diag::err_drv_unsupported_opt_for_target)
+ << A->getAsString(Args) << TripleStr;
+ }
+ }
+
if (Arg *A = Args.getLastArg(options::OPT_fbasic_block_sections_EQ)) {
StringRef Val = A->getValue();
if (Triple.isX86() && Triple.isOSBinFormatELF()) {
diff --git a/clang/test/Driver/basic-block-address-map.c b/clang/test/Driver/basic-block-address-map.c
new file mode 100644
index 0000000000000..022f972b412d6
--- /dev/null
+++ b/clang/test/Driver/basic-block-address-map.c
@@ -0,0 +1,8 @@
+// RUN: %clang -### -target x86_64 -fbasic-block-address-map %s -S 2>&1 | FileCheck -check-prefix=CHECK-PRESENT %s
+// CHECK-PRESENT: -fbasic-block-address-map
+
+// RUN: %clang -### -target x86_64 -fno-basic-block-address-map %s -S 2>&1 | FileCheck %s --check-prefix=CHECK-ABSENT
+// CHECK-ABSENT-NOT: -fbasic-block-address-map
+
+// RUN: not %clang -c -target x86_64-apple-darwin10 -fbasic-block-address-map %s -S 2>&1 | FileCheck -check-prefix=CHECK-TRIPLE %s
+// CHECK-TRIPLE: error: unsupported option '-fbasic-block-address-map' for target
diff --git a/lld/ELF/Config.h b/lld/ELF/Config.h
index 3a4a47d900fe7..fcca8c42b29b7 100644
--- a/lld/ELF/Config.h
+++ b/lld/ELF/Config.h
@@ -187,6 +187,7 @@ struct Config {
llvm::StringRef cmseOutputLib;
StringRef zBtiReport = "none";
StringRef zCetReport = "none";
+ bool ltoBBAddrMap;
llvm::StringRef ltoBasicBlockSections;
std::pair<llvm::StringRef, llvm::StringRef> thinLTOObjectSuffixReplace;
llvm::StringRef thinLTOPrefixReplaceOld;
diff --git a/lld/ELF/Driver.cpp b/lld/ELF/Driver.cpp
index f4b7d1c9d5b97..c19fba6598a8a 100644
--- a/lld/ELF/Driver.cpp
+++ b/lld/ELF/Driver.cpp
@@ -1298,6 +1298,9 @@ static void readConfigs(opt::InputArgList &args) {
config->ltoObjPath = args.getLastArgValue(OPT_lto_obj_path_eq);
config->ltoPartitions = args::getInteger(args, OPT_lto_partitions, 1);
config->ltoSampleProfile = args.getLastArgValue(OPT_lto_sample_profile);
+ config->ltoBBAddrMap =
+ args.hasFlag(OPT_lto_basic_block_address_map,
+ OPT_no_lto_basic_block_address_map, false);
config->ltoBasicBlockSections =
args.getLastArgValue(OPT_lto_basic_block_sections);
config->ltoUniqueBasicBlockSectionNames =
diff --git a/lld/ELF/LTO.cpp b/lld/ELF/LTO.cpp
index c39c6e6ea74ba..3d92007469263 100644
--- a/lld/ELF/LTO.cpp
+++ b/lld/ELF/LTO.cpp
@@ -61,6 +61,8 @@ static lto::Config createConfig() {
c.Options.FunctionSections = true;
c.Options.DataSections = true;
+ c.Options.BBAddrMap = config->ltoBBAddrMap;
+
// Check if basic block sections must be used.
// Allowed values for --lto-basic-block-sections are "all", "labels",
// "<file name specifying basic block ids>", or none. This is the equivalent
diff --git a/lld/ELF/Options.td b/lld/ELF/Options.td
index c2c9cabc92a4d..c10a73e2d9c36 100644
--- a/lld/ELF/Options.td
+++ b/lld/ELF/Options.td
@@ -637,6 +637,9 @@ def save_temps_eq: JJ<"save-temps=">, HelpText<"Save select intermediate LTO com
Values<"resolution,preopt,promote,internalize,import,opt,precodegen,prelink,combinedindex">;
def lto_basic_block_sections: JJ<"lto-basic-block-sections=">,
HelpText<"Enable basic block sections for LTO">;
+defm lto_basic_block_address_map: BB<"lto-basic-block-address-map",
+ "Emit basic block address map for LTO",
+ "Do not emit basic block address map for LTO (default)">;
defm lto_unique_basic_block_section_names: BB<"lto-unique-basic-block-section-names",
"Give unique names to every basic block section for LTO",
"Do not give unique names to every basic block section for LTO (default)">;
diff --git a/lld/test/ELF/lto/basic-block-address-map.ll b/lld/test/ELF/lto/basic-block-address-map.ll
new file mode 100644
index 0000000000000..b96e7a8401f85
--- /dev/null
+++ b/lld/test/ELF/lto/basic-block-address-map.ll
@@ -0,0 +1,28 @@
+; REQUIRES: x86
+; RUN: llvm-as %s -o %t.o
+; RUN: ld.lld %t.o -o %t --lto-basic-block-address-map --lto-O0 --save-temps
+; RUN: llvm-readobj --sections %t.lto.o | FileCheck --check-prefix=SECNAMES %s
+
+; SECNAMES: Type: SHT_LLVM_BB_ADDR_MAP
+
+target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-f80:128-n8:16:32:64-S128"
+target triple = "x86_64-unknown-linux-gnu"
+
+; Function Attrs: nounwind uwtable
+define dso_local void @foo(i32 %b) local_unnamed_addr {
+entry:
+ %tobool.not = icmp eq i32 %b, 0
+ br i1 %tobool.not, label %if.end, label %if.then
+
+if.then: ; preds = %entry
+ tail call void @foo(i32 0)
+ br label %if.end
+
+if.end: ; preds = %entry, %if.then
+ ret void
+}
+
+define void @_start() {
+ call void @foo(i32 1)
+ ret void
+}
diff --git a/llvm/include/llvm/CodeGen/CommandFlags.h b/llvm/include/llvm/CodeGen/CommandFlags.h
index bf166a63edb07..d74d724500607 100644
--- a/llvm/include/llvm/CodeGen/CommandFlags.h
+++ b/llvm/include/llvm/CodeGen/CommandFlags.h
@@ -159,6 +159,8 @@ struct RegisterCodeGenFlags {
RegisterCodeGenFlags();
};
+bool getEnableBBAddrMap();
+
llvm::BasicBlockSection getBBSectionsMode(llvm::TargetOptions &Options);
/// Common utility function tightly tied to the options listed here. Initializes
diff --git a/llvm/include/llvm/Object/ELFTypes.h b/llvm/include/llvm/Object/ELFTypes.h
index 956f7811dd6c3..810f2156e5f5a 100644
--- a/llvm/include/llvm/Object/ELFTypes.h
+++ b/llvm/include/llvm/Object/ELFTypes.h
@@ -796,6 +796,47 @@ template <class ELFT> struct Elf_Mips_ABIFlags {
// Struct representing the BBAddrMap for one function.
struct BBAddrMap {
+
+ // Bitfield of optional features to control the extra information
+ // emitted/encoded in the the section.
+ struct Features {
+ bool FuncEntryCount : 1;
+ bool BBFreq : 1;
+ bool BrProb : 1;
+ bool MultiBBRange : 1;
+
+ bool hasPGOAnalysis() const { return FuncEntryCount || BBFreq || BrProb; }
+
+ bool hasPGOAnalysisBBData() const { return BBFreq || BrProb; }
+
+ // Encodes to minimum bit width representation.
+ uint8_t encode() const {
+ return (static_cast<uint8_t>(FuncEntryCount) << 0) |
+ (static_cast<uint8_t>(BBFreq) << 1) |
+ (static_cast<uint8_t>(BrProb) << 2) |
+ (static_cast<uint8_t>(MultiBBRange) << 3);
+ }
+
+ // Decodes from minimum bit width representation and validates no
+ // unnecessary bits are used.
+ static Expected<Features> decode(uint8_t Val) {
+ Features Feat{
+ static_cast<bool>(Val & (1 << 0)), static_cast<bool>(Val & (1 << 1)),
+ static_cast<bool>(Val & (1 << 2)), static_cast<bool>(Val & (1 << 3))};
+ if (Feat.encode() != Val)
+ return createStringError(
+ std::error_code(), "invalid encoding for BBAddrMap::Features: 0x%x",
+ Val);
+ return Feat;
+ }
+
+ bool operator==(const Features &Other) const {
+ return std::tie(FuncEntryCount, BBFreq, BrProb, MultiBBRange) ==
+ std::tie(Other.FuncEntryCount, Other.BBFreq, Other.BrProb,
+ Other.MultiBBRange);
+ }
+ };
+
// Struct representing the BBAddrMap information for one basic block.
struct BBEntry {
struct Metadata {
@@ -839,7 +880,7 @@ struct BBAddrMap {
};
uint32_t ID; // Unique ID of this basic block.
- uint32_t Offset; // Offset of basic block relative to function start.
+ uint32_t Offset; // Offset of basic block relative to the base address.
uint32_t Size; // Size of the basic block.
Metadata MD; // Metdata for this basic block.
@@ -858,62 +899,63 @@ struct BBAddrMap {
bool hasIndirectBranch() const { return MD.HasIndirectBranch; }
};
- BBAddrMap(uint64_t Addr, std::vector<BBEntry> BBEntries)
- : Addr(Addr), BBEntries(std::move(BBEntries)) {}
+ // Struct representing the BBAddrMap information for a contiguous range of
+ // basic blocks (a function or a basic block section).
+ struct BBRangeEntry {
+ uint64_t BaseAddress; // Base address of the range.
+ std::vector<BBEntry> BBEntries; // Basic block entries for this range.
+
+ // Equality operator for unit testing.
+ bool operator==(const BBRangeEntry &Other) const {
+ return BaseAddress == Other.BaseAddress &&
+ std::equal(BBEntries.begin(), BBEntries.end(),
+ Other.BBEntries.begin());
+ }
+ };
- // Returns the address of the corresponding function.
- uint64_t getFunctionAddress() const { return Addr; }
+ // All ranges for this function. The first range always corresponds to the
+ // function entry.
+ std::vector<BBRangeEntry> BBRanges;
- // Returns the basic block entries for this function.
- const std::vector<BBEntry> &getBBEntries() const { return BBEntries; }
+ // Returns the function address associated with this BBAddrMap, which is
+ // stored as the `BaseAddress` of its first BBRangeEntry. Returns 0 if
+ // BBRanges is empty.
+ uint64_t getFunctionAddress() const {
+ if (BBRanges.empty())
+ return 0;
+ return BBRanges.front().BaseAddress;
+ }
+
+ // Returns the total number of bb entries in all bb ranges.
+ size_t getNumBBEntries() const {
+ size_t NumBBEntries = 0;
+ for (const auto &BBR: BBRanges)
+ NumBBEntries += BBR.BBEntries.size();
+ return NumBBEntries;
+ }
+
+ // Returns the index of the bb range with the given base address, or
+ // `std::nullopt` if no such range exists.
+ std::optional<size_t> getBBRangeIndexForBaseAddress(uint64_t BaseAddress) const {
+ for (size_t I=0; I<BBRanges.size(); ++I)
+ if (BBRanges[I].BaseAddress == BaseAddress)
+ return I;
+ return {};
+ }
+
+ // Returns bb entries in the bb range indexed `BBRangeIndex`.
+ const std::vector<BBEntry> &getBBEntries(unsigned BBRangeIndex = 0) const {
+ return BBRanges.at(BBRangeIndex).BBEntries;
+ }
// Equality operator for unit testing.
bool operator==(const BBAddrMap &Other) const {
- return Addr == Other.Addr && std::equal(BBEntries.begin(), BBEntries.end(),
- Other.BBEntries.begin());
+ return std::equal(BBRanges.begin(), BBRanges.end(), Other.BBRanges.begin());
}
-
- uint64_t Addr; // Function address
- std::vector<BBEntry> BBEntries; // Basic block entries for this function.
};
/// A feature extension of BBAddrMap that holds information relevant to PGO.
struct PGOAnalysisMap {
- /// Bitfield of optional features to include in the PGO extended map.
- struct Features {
- bool FuncEntryCount : 1;
- bool BBFreq : 1;
- bool BrProb : 1;
-
- // True if at least one feature is enabled
- bool anyEnabled() const { return FuncEntryCount || BBFreq || BrProb; }
-
- // Encodes to minimum bit width representation.
- uint8_t encode() const {
- return (static_cast<uint8_t>(FuncEntryCount) << 0) |
- (static_cast<uint8_t>(BBFreq) << 1) |
- (static_cast<uint8_t>(BrProb) << 2);
- }
-
- // Decodes from minimum bit width representation and validates no
- // unnecessary bits are used.
- static Expected<Features> decode(uint8_t Val) {
- Features Feat{static_cast<bool>(Val & (1 << 0)),
- static_cast<bool>(Val & (1 << 1)),
- static_cast<bool>(Val & (1 << 2))};
- if (Feat.encode() != Val)
- return createStringError(
- std::error_code(),
- "invalid encoding for PGOAnalysisMap::Features: 0x%x", Val);
- return Feat;
- }
-
- bool operator==(const Features &Other) const {
- return std::tie(FuncEntryCount, BBFreq, BrProb) ==
- std::tie(Other.FuncEntryCount, Other.BBFreq, Other.BrProb);
- }
- };
-
/// Extra basic block data with fields for block frequency and branch
/// probability.
struct PGOBBEntry {
@@ -945,7 +987,7 @@ struct PGOAnalysisMap {
std::vector<PGOBBEntry> BBEntries; // Extended basic block entries
// Flags to indicate if each PGO related info was enabled in this function
- Features FeatEnable;
+ BBAddrMap::Features FeatEnable;
bool operator==(const PGOAnalysisMap &Other) const {
return std::tie(FuncEntryCount, BBEntries, FeatEnable) ==
diff --git a/llvm/include/llvm/ObjectYAML/ELFYAML.h b/llvm/include/llvm/ObjectYAML/ELFYAML.h
index 12b47c271da2c..8f045d6383623 100644
--- a/llvm/include/llvm/ObjectYAML/ELFYAML.h
+++ b/llvm/include/llvm/ObjectYAML/ELFYAML.h
@@ -165,9 +165,21 @@ struct BBAddrMapEntry {
};
uint8_t Version;
llvm::yaml::Hex8 Feature;
- llvm::yaml::Hex64 Address;
- std::optional<uint64_t> NumBlocks;
- std::optional<std::vector<BBEntry>> BBEntries;
+
+ struct BBRangeEntry {
+ llvm::yaml::Hex64 BaseAddress;
+ std::optional<uint64_t> NumBlocks;
+ std::optional<std::vector<BBEntry>> BBEntries;
+ };
+
+ std::optional<uint64_t> NumBBRanges;
+ std::optional<std::vector<BBRangeEntry>> BBRanges;
+
+ llvm::yaml::Hex64 getFunctionAddress() const {
+ if (!BBRanges || BBRanges->empty())
+ return 0;
+ return BBRanges->front().BaseAddress;
+ }
};
struct PGOAnalysisMapEntry {
@@ -751,6 +763,7 @@ bool shouldAllocateFileSpace(ArrayRef<ProgramHeader> Phdrs,
LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::StackSizeEntry)
LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::BBAddrMapEntry)
LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::BBAddrMapEntry::BBEntry)
+LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::BBAddrMapEntry::BBRangeEntry)
LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::PGOAnalysisMapEntry)
LLVM_YAML_IS_SEQUENCE_VECTOR(llvm::ELFYAML::PGOAnalysisMapEntry::PGOBBEntry)
LLVM_YAML_IS_SEQUENCE_VECTOR(
@@ -916,11 +929,15 @@ template <> struct MappingTraits<ELFYAML::StackSizeEntry> {
};
template <> struct MappingTraits<ELFYAML::BBAddrMapEntry> {
- static void mapping(IO &IO, ELFYAML::BBAddrMapEntry &Rel);
+ static void mapping(IO &IO, ELFYAML::BBAddrMapEntry &E);
+};
+
+template <> struct MappingTraits<ELFYAML::BBAddrMapEntry::BBRangeEntry> {
+ static void mapping(IO &IO, ELFYAML::BBAddrMapEntry::BBRangeEntry &E);
};
template <> struct MappingTraits<ELFYAML::BBAddrMapEntry::BBEntry> {
- static void mapping(IO &IO, ELFYAML::BBAddrMapEntry::BBEntry &Rel);
+ static void mapping(IO &IO, ELFYAML::BBAddrMapEntry::BBEntry &E);
};
template <> struct MappingTraits<ELFYAML::PGOAnalysisMapEntry> {
diff --git a/llvm/include/llvm/Target/TargetOptions.h b/llvm/include/llvm/Target/TargetOptions.h
index d02d1699813c9..4bc03b0dc80a4 100644
--- a/llvm/include/llvm/Target/TargetOptions.h
+++ b/llvm/include/llvm/Target/TargetOptions.h
@@ -148,7 +148,7 @@ namespace llvm {
EmulatedTLS(false), EnableTLSDESC(false), EnableIPRA(false),
EmitStackSizeSection(false), EnableMachineOutliner(false),
EnableMachineFunctionSplitter(false), SupportsDefaultOutlining(false),
- EmitAddrsig(false), EmitCallSiteInfo(false),
+ EmitAddrsig(false), BBAddrMap(false), EmitCallSiteInfo(false),
SupportsDebugEntryValues(false), EnableDebugEntryValues(false),
ValueTrackingVariableLocations(false), ForceDwarfFrameSection(false),
XRayFunctionIndex(true), DebugStrictDwarf(false), Hotpatch(false),
@@ -295,9 +295,6 @@ namespace llvm {
/// function in the runtime library..
unsigned EmulatedTLS : 1;
- /// EnableTLSDESC - This flag enables TLS Descriptors.
- unsigned EnableTLSDESC : 1;
-
/// This flag enables InterProcedural Register Allocation (IPRA).
unsigned EnableIPRA : 1;
diff --git a/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp b/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
index 0d573562de969..b961fc2c338ae 100644
--- a/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
+++ b/llvm/lib/CodeGen/AsmPrinter/AsmPrinter.cpp
@@ -151,8 +151,9 @@ static cl::bits<PGOMapFeaturesEnum> PgoAnalysisMapFeatures(
"Basic Block Frequency"),
clEnumValN(PGOMapFeaturesEnum::BrProb, "br-prob",
"Branch Probability")),
- cl::desc("Enable extended information within the BBAddrMap that is "
- "extracted from PGO related analysis."));
+ cl::desc(
+ "Enable extended information within the SHT_LLVM_BB_ADDR_MAP that is "
+ "extracted from PGO related analysis."));
const char DWARFGroupName[] = "dwarf";
const char DWARFGroupDescription[] = "DWARF Emission";
@@ -1371,6 +1372,14 @@ static uint32_t getBBAddrMapMetadata(const MachineBasicBlock &MBB) {
.encode();
}
+static llvm::object::BBAddrMap::Features
+getBBAddrMapFeature(const MachineFunction &MF, int NumMBBSectionRanges) {
+ return {PgoAnalysisMapFeatures.isSet(PGOMapFeaturesEnum::FuncEntryCount),
+ PgoAnalysisMapFeatures.isSet(PGOMapFeaturesEnum::BBFreq),
+ PgoAnalysisMapFeatures.isSet(PGOMapFeaturesEnum::BrProb),
+ MF.hasBBSections() && NumMBBSectionRanges > 1};
+}
+
void AsmPrinter::emitBBAddrMapSection(const MachineFunction &MF) {
MCSection *BBAddrMapSection =
getObjFileLowering().getBBAddrMapSection(*MF.getSection());
@@ -1384,17 +1393,48 @@ void AsmPrinter::emitBBAddrMapSection(const MachineFunction &MF) {
uint8_t BBAddrMapVersion = OutStreamer->getContext().getBBAddrMapVersion();
OutStreamer->emitInt8(BBAddrMapVersion);
OutStreamer->AddComment("feature");
- auto FeaturesBits = static_cast<uint8_t>(PgoAnalysisMapFeatures.getBits());
- OutStreamer->emitInt8(FeaturesBits);
- OutStreamer->AddComment("function address");
- OutStreamer->emitSymbolValue(FunctionSymbol, getPointerSize());
- OutStreamer->AddComment("number of basic blocks");
- OutStreamer->emitULEB128IntValue(MF.size());
- const MCSymbol *PrevMBBEndSymbol = FunctionSymbol;
+ auto Features = getBBAddrMapFeature(MF, MBBSectionRanges.size());
+ OutStreamer->emitInt8(Features.encode());
// Emit BB Information for each basic block in the function.
+ if (Features.MultiBBRange) {
+ OutStreamer->AddComment("number of basic block ranges");
+ OutStreamer->emitULEB128IntValue(MBBSectionRanges.size());
+ }
+ // Number of blocks in each MBB section.
+ MapVector<unsigned, unsigned> MBBSectionNumBlocks;
+ const MCSymbol *PrevMBBEndSymbol = nullptr;
+ if (!Features.MultiBBRange) {
+ OutStreamer->AddComment("function address");
+ OutStreamer->emitSymbolValue(FunctionSymbol, getPointerSize());
+ OutStreamer->AddComment("number of basic blocks");
+ OutStreamer->emitULEB128IntValue(MF.size());
+ PrevMBBEndSymbol = FunctionSymbol;
+ } else {
+ unsigned BBCount = 0;
+ for (const MachineBasicBlock &MBB : MF) {
+ BBCount++;
+ if (MBB.isEndSection()) {
+ // Store each section's basic block count when it ends.
+ MBBSectionNumBlocks[MBB.getSectionIDNum()] = BBCount;
+ // Reset the count for the next section.
+ BBCount = 0;
+ }
+ }
+ }
+ // Emit the BB entry for each basic block in the function.
for (const MachineBasicBlock &MBB : MF) {
const MCSymbol *MBBSymbol =
MBB.isEntryBlock() ? FunctionSymbol : MBB.getSymbol();
+ bool IsBeginSection =
+ Features.MultiBBRange && (MBB.isBeginSection() || MBB.isEntryBlock());
+ if (IsBeginSection) {
+ OutStreamer->AddComment("base address");
+ OutStreamer->emitSymbolValue(MBBSymbol, getPointerSize());
+ OutStreamer->AddComment("number of basic blocks");
+ OutStreamer->emitULEB128IntValue(
+ MBBSectionNumBlocks[MBB.getSectionIDNum()]);
+ PrevMBBEndSymbol = MBBSymbol;
+ }
// TODO: Remove this check when version 1 is deprecated.
if (BBAddrMapVersion > 1) {
OutStreamer->AddComment("BB id");
@@ -1416,35 +1456,32 @@ void AsmPrinter::emitBBAddrMapSection(const MachineFunction &MF) {
PrevMBBEndSymbol = MBB.getEndSymbol();
}
- if (FeaturesBits != 0) {
+ if (Features.hasPGOAnalysis()) {
assert(BBAddrMapVersion >= 2 &&
"PGOAnalysisMap only supports version 2 or later");
- auto FeatEnable =
- cantFail(object::PGOAnalysisMap::Features::decode(FeaturesBits));
-
- if (FeatEnable.FuncEntryCount) {
+ if (Features.FuncEntryCount) {
OutStreamer->AddComment("function entry count");
auto MaybeEntryCount = MF.getFunction().getEntryCount();
OutStreamer->emitULEB128IntValue(
MaybeEntryCount ? MaybeEntryCount->getCount() : 0);
}
const MachineBlockFrequencyInfo *MBFI =
- FeatEnable.BBFreq
+ Features.BBFreq
? &getAnalysis<LazyMachineBlockFrequencyInfoPass>().getBFI()
: nullptr;
const MachineBranchProbabilityInfo *MBPI =
- FeatEnable.BrProb ? &getAnalysis<MachineBranchProbabilityInfo>()
- : nullptr;
+ Features.BrProb ? &getAnalysis<MachineBranchProbabilityInfo>()
+ : nullptr;
- if (FeatEnable.BBFreq || FeatEnable.BrProb) {
+ if (Features.BBFreq || Features.BrProb) {
for (const MachineBasicBlock &MBB : MF) {
- if (FeatEnable.BBFreq) {
+ if (Features.BBFreq) {
OutStreamer->AddComment("basic block frequency");
OutStreamer->emitULEB128IntValue(
MBFI->getBlockFreq(&MBB).getFrequency());
}
- if (FeatEnable.BrProb) {
+ if (Features.BrProb) {
unsigned SuccCount = MBB.succ_size();
OutStreamer->AddComment("basic block successor count");
OutStreamer->emitULEB128IntValue(SuccCount);
@@ -1833,7 +1870,7 @@ void AsmPrinter::emitFunctionBody() {
// We must emit temporary symbol for the end of this basic block, if either
// we have BBLabels enabled or if this basic blocks marks the end of a
// section.
- if (MF->hasBBLabels() ||
+ if (MF->hasBBLabels() || MF->getTarget().Options.BBAddrMap ||
(MAI->hasDotTypeDotSizeDirective() && MBB.isEndSection()))
OutStreamer->emitLabel(MBB.getEndSymbol());
@@ -1986,7 +2023,7 @@ void AsmPrinter::emitFunctionBody() {
// Emit section containing BB address offsets and their metadata, when
// BB labels are requested for this function. Skip empty functions.
if (HasAnyRealCode) {
- if (MF->hasBBLabels())
+ if (MF->hasBBLabels() || MF->getTarget().Options.BBAddrMap)
emitBBAddrMapSection(*MF);
else if (PgoAnalysisMapFeatures.getBits() != 0)
MF->getContext().reportWarning(
@@ -2580,9 +2617,9 @@ void AsmPrinter::SetupMachineFunction(MachineFunction &MF) {
bool NeedsLocalForSize = MAI->needsLocalForSize();
if (F.hasFnAttribute("patchable-function-entry") ||
F.hasFnAttribute("function-instrument") ||
- F.hasFnAttribute("xray-instruction-threshold") ||
- needFuncLabels(MF) || NeedsLocalForSize ||
- MF.getTarget().Options.EmitStackSizeSection || MF.hasBBLabels()) {
+ F.hasFnAttribute("xray-instruction-threshold") || needFuncLabels(MF) ||
+ NeedsLocalForSize || MF.getTarget().Options.EmitStackSizeSection ||
+ MF.getTarget().Options.BBAddrMap || MF.hasBBLabels()) {
CurrentFnBegin = createTempSymbol("func_begin");
if (NeedsLocalForSize)
CurrentFnSymForSize = CurrentFnBegin;
@@ -4049,7 +4086,9 @@ bool AsmPrinter::shouldEmitLabelForBasicBlock(
// With `-fbasic-block-sections=`, a label is needed for every non-entry block
// in the labels mode (option `=labels`) and every section beginning in the
// sections mode (`=all` and `=list=`).
- if ((MF->hasBBLabels() || MBB.isBeginSection()) && !MBB.isEntryBlock())
+ if ((MF->hasBBLabels() || MF->getTarget().Options.BBAddrMap ||
+ MBB.isBeginSection()) &&
+ !MBB.isEntryBlock())
return true;
// A label is needed for any block with at least one predecessor (when that
// predecessor is not the fallthrough predecessor, or if it is an EH funclet
diff --git a/llvm/lib/CodeGen/BasicBlockSections.cpp b/llvm/lib/CodeGen/BasicBlockSections.cpp
index dbb6ebb3d7ebe..eb3f9e7078f1a 100644
--- a/llvm/lib/CodeGen/BasicBlockSections.cpp
+++ b/llvm/lib/CodeGen/BasicBlockSections.cpp
@@ -57,10 +57,10 @@
// function into potentially several disjoint pieces, and CFI needs to be
// emitted per cluster. This also bloats the object file and binary sizes.
//
-// Basic Block Labels
+// Basic Block Address Map
// ==================
//
-// With -fbasic-block-sections=labels, we encode the offsets of BB addresses of
+// With -fbasic-block-address-map, we emit the offsets of BB addresses of
// every function into the .llvm_bb_addr_map section. Along with the function
// symbols, this allows for mapping of virtual addresses in PMU profiles back to
// the corresponding basic blocks. This logic is implemented in AsmPrinter. This
@@ -118,6 +118,10 @@ class BasicBlockSections : public MachineFunctionPass {
/// Identify basic blocks that need separate sections and prepare to emit them
/// accordingly.
bool runOnMachineFunction(MachineFunction &MF) override;
+
+private:
+ bool handleBBSections(MachineFunction &MF);
+ bool handleBBAddrMap(MachineFunction &MF);
};
} // end anonymous namespace
@@ -280,10 +284,12 @@ bool llvm::hasInstrProfHashMismatch(MachineFunction &MF) {
return false;
}
-bool BasicBlockSections::runOnMachineFunction(MachineFunction &MF) {
+// Identify, arrange, and modify basic blocks which need separate sections
+// according to the specification provided by the -fbasic-block-sections flag.
+bool BasicBlockSections::handleBBSections(MachineFunction &MF) {
auto BBSectionsType = MF.getTarget().getBBSectionsType();
- assert(BBSectionsType != BasicBlockSection::None &&
- "BB Sections not enabled!");
+ if (BBSectionsType == BasicBlockSection::None)
+ return false;
// Check for source drift. If the source has changed since the profiles
// were obtained, optimizing basic blocks might be sub-optimal.
@@ -300,7 +306,7 @@ bool BasicBlockSections::runOnMachineFunction(MachineFunction &MF) {
if (BBSectionsType == BasicBlockSection::Labels) {
MF.setBBSectionsType(BBSectionsType);
- return false;
+ return true;
}
DenseMap<UniqueBBID, BBClusterInfo> FuncClusterInfo;
@@ -364,6 +370,27 @@ bool BasicBlockSections::runOnMachineFunction(MachineFunction &MF) {
return true;
}
+// When the BB address map needs to be generated, this renumbers basic blocks to
+// make them appear in increasing order of their IDs in the function. This
+// avoids the need to store basic block IDs in the BB address map section, since
+// they can be determined implicitly.
+bool BasicBlockSections::handleBBAddrMap(MachineFunction &MF) {
+ if (MF.getTarget().getBBSectionsType() == BasicBlockSection::Labels)
+ return false;
+ if (!MF.getTarget().Options.BBAddrMap)
+ return false;
+ MF.RenumberBlocks();
+ return true;
+}
+
+bool BasicBlockSections::runOnMachineFunction(MachineFunction &MF) {
+ // First handle the basic block sections.
+ auto R1 = handleBBSections(MF);
+ // Handle basic block address map after basic block sections are finalized.
+ auto R2 = handleBBAddrMap(MF);
+ return R1 || R2;
+}
+
void BasicBlockSections::getAnalysisUsage(AnalysisUsage &AU) const {
AU.setPreservesAll();
AU.addRequired<BasicBlockSectionsProfileReaderWrapperPass>();
diff --git a/llvm/lib/CodeGen/CommandFlags.cpp b/llvm/lib/CodeGen/CommandFlags.cpp
index 51406fb287e66..d82b2a85016c2 100644
--- a/llvm/lib/CodeGen/CommandFlags.cpp
+++ b/llvm/lib/CodeGen/CommandFlags.cpp
@@ -90,6 +90,7 @@ CGOPT_EXP(bool, DataSections)
CGOPT_EXP(bool, FunctionSections)
CGOPT(bool, IgnoreXCOFFVisibility)
CGOPT(bool, XCOFFTracebackTable)
+CGOPT(bool, EnableBBAddrMap)
CGOPT(std::string, BBSections)
CGOPT(unsigned, TLSSize)
CGOPT_EXP(bool, EmulatedTLS)
@@ -390,6 +391,11 @@ codegen::RegisterCodeGenFlags::RegisterCodeGenFlags() {
cl::init(true));
CGBINDOPT(XCOFFTracebackTable);
+ static cl::opt<bool> EnableBBAddrMap(
+ "basic-block-address-map",
+ cl::desc("Emit the basic block address map section"), cl::init(false));
+ CGBINDOPT(EnableBBAddrMap);
+
static cl::opt<std::string> BBSections(
"basic-block-sections",
cl::desc("Emit basic blocks into separate sections"),
@@ -568,6 +574,7 @@ codegen::InitTargetOptionsFromCodeGenFlags(const Triple &TheTriple) {
Options.FunctionSections = getFunctionSections();
Options.IgnoreXCOFFVisibility = getIgnoreXCOFFVisibility();
Options.XCOFFTracebackTable = getXCOFFTracebackTable();
+ Options.BBAddrMap = getEnableBBAddrMap();
Options.BBSections = getBBSectionsMode(Options);
Options.UniqueSectionNames = getUniqueSectionNames();
Options.UniqueBasicBlockSectionNames = getUniqueBasicBlockSectionNames();
diff --git a/llvm/lib/CodeGen/MachineFunction.cpp b/llvm/lib/CodeGen/MachineFunction.cpp
index 57af571ed9bfd..323f1a6a1b2bd 100644
--- a/llvm/lib/CodeGen/MachineFunction.cpp
+++ b/llvm/lib/CodeGen/MachineFunction.cpp
@@ -467,6 +467,7 @@ MachineFunction::CreateMachineBasicBlock(const BasicBlock *BB,
// `-basic-block-sections=list` to allow robust mapping of profiles to basic
// blocks.
if (Target.getBBSectionsType() == BasicBlockSection::Labels ||
+ Target.Options.BBAddrMap ||
Target.getBBSectionsType() == BasicBlockSection::List)
MBB->setBBID(BBID.has_value() ? *BBID : UniqueBBID{NextBBID++, 0});
return MBB;
diff --git a/llvm/lib/CodeGen/TargetPassConfig.cpp b/llvm/lib/CodeGen/TargetPassConfig.cpp
index 599ca48189048..e82f14e878141 100644
--- a/llvm/lib/CodeGen/TargetPassConfig.cpp
+++ b/llvm/lib/CodeGen/TargetPassConfig.cpp
@@ -1221,19 +1221,13 @@ void TargetPassConfig::addMachinePasses() {
addPass(createMIRAddFSDiscriminatorsPass(
sampleprof::FSDiscriminatorPass::PassLast));
+ bool NeedsBBSections =
+ TM->getBBSectionsType() != llvm::BasicBlockSection::None;
// Machine function splitter uses the basic block sections feature. Both
- // cannot be enabled at the same time. Basic block sections takes precedence.
- // FIXME: In principle, BasicBlockSection::Labels and splitting can used
- // together. Update this check once we have addressed any issues.
- if (TM->getBBSectionsType() != llvm::BasicBlockSection::None) {
- if (TM->getBBSectionsType() == llvm::BasicBlockSection::List) {
- addPass(llvm::createBasicBlockSectionsProfileReaderWrapperPass(
- TM->getBBSectionsFuncListBuf()));
- addPass(llvm::createBasicBlockPathCloningPass());
- }
- addPass(llvm::createBasicBlockSectionsPass());
- } else if (TM->Options.EnableMachineFunctionSplitter ||
- EnableMachineFunctionSplitter) {
+ // cannot be enabled at the same time. We do not apply machine function
+ // splitter if -basic-block-sections is requested.
+ if (!NeedsBBSections && (TM->Options.EnableMachineFunctionSplitter ||
+ EnableMachineFunctionSplitter)) {
const std::string ProfileFile = getFSProfileFile(TM);
if (!ProfileFile.empty()) {
if (EnableFSDiscriminator) {
@@ -1250,6 +1244,16 @@ void TargetPassConfig::addMachinePasses() {
}
addPass(createMachineFunctionSplitterPass());
}
+ // We run the BasicBlockSections pass if either we need BB sections or BB
+ // address map (or both).
+ if (NeedsBBSections || TM->Options.BBAddrMap) {
+ if (TM->getBBSectionsType() == llvm::BasicBlockSection::List) {
+ addPass(llvm::createBasicBlockSectionsProfileReaderWrapperPass(
+ TM->getBBSectionsFuncListBuf()));
+ addPass(llvm::createBasicBlockPathCloningPass());
+ }
+ addPass(llvm::createBasicBlockSectionsPass());
+ }
addPostBBSections();
diff --git a/llvm/lib/Object/ELF.cpp b/llvm/lib/Object/ELF.cpp
index f24395b02043d..d7f75a6303673 100644
--- a/llvm/lib/Object/ELF.cpp
+++ b/llvm/lib/Object/ELF.cpp
@@ -693,6 +693,17 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
for (typename ELFFile<ELFT>::Elf_Rela Rela : *Relas)
FunctionOffsetTranslations[Rela.r_offset] = Rela.r_addend;
}
+ auto GetAddressForRelocation =
+ [&](unsigned RelocationOffsetInSection) -> Expected<unsigned> {
+ auto FOTIterator =
+ FunctionOffsetTranslations.find(RelocationOffsetInSection);
+ if (FOTIterator == FunctionOffsetTranslations.end()) {
+ return createError("failed to get relocation data for offset: " +
+ Twine::utohexstr(RelocationOffsetInSection) +
+ " in section " + describe(EF, Sec));
+ }
+ return FOTIterator->second;
+ };
Expected<ArrayRef<uint8_t>> ContentsOrErr = EF.getSectionContents(Sec);
if (!ContentsOrErr)
return ContentsOrErr.takeError();
@@ -704,9 +715,26 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
Error ULEBSizeErr = Error::success();
Error MetadataDecodeErr = Error::success();
+ // Helper lampda to extract the (possiblly relocatable) address stored at Cur.
+ auto ExtractAddress = [&]() -> Expected<typename ELFFile<ELFT>::uintX_t> {
+ uint64_t RelocationOffsetInSection = Cur.tell();
+ auto Address =
+ static_cast<typename ELFFile<ELFT>::uintX_t>(Data.getAddress(Cur));
+ if (!Cur)
+ return Cur.takeError();
+ if (!IsRelocatable)
+ return Address;
+ assert(Address == 0);
+ Expected<unsigned> AddressOrErr =
+ GetAddressForRelocation(RelocationOffsetInSection);
+ if (!AddressOrErr)
+ return AddressOrErr.takeError();
+ return *AddressOrErr;
+ };
+
uint8_t Version = 0;
uint8_t Feature = 0;
- PGOAnalysisMap::Features FeatEnable{};
+ BBAddrMap::Features FeatEnable{};
while (!ULEBSizeErr && !MetadataDecodeErr && Cur &&
Cur.tell() < Content.size()) {
if (Sec.sh_type == ELF::SHT_LLVM_BB_ADDR_MAP) {
@@ -719,11 +747,10 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
Feature = Data.getU8(Cur); // Feature byte
if (!Cur)
break;
- auto FeatEnableOrErr = PGOAnalysisMap::Features::decode(Feature);
+ auto FeatEnableOrErr = BBAddrMap::Features::decode(Feature);
if (!FeatEnableOrErr)
return FeatEnableOrErr.takeError();
- FeatEnable =
- FeatEnableOrErr ? *FeatEnableOrErr : PGOAnalysisMap::Features{};
+ FeatEnable = *FeatEnableOrErr;
if (Feature != 0 && Version < 2 && Cur)
return createError(
"version should be >= 2 for SHT_LLVM_BB_ADDR_MAP when "
@@ -731,50 +758,59 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
Twine(static_cast<int>(Version)) +
" feature = " + Twine(static_cast<int>(Feature)));
}
- uint64_t SectionOffset = Cur.tell();
- auto Address =
- static_cast<typename ELFFile<ELFT>::uintX_t>(Data.getAddress(Cur));
- if (!Cur)
- return Cur.takeError();
- if (IsRelocatable) {
- assert(Address == 0);
- auto FOTIterator = FunctionOffsetTranslations.find(SectionOffset);
- if (FOTIterator == FunctionOffsetTranslations.end()) {
- return createError("failed to get relocation data for offset: " +
- Twine::utohexstr(SectionOffset) + " in section " +
- describe(EF, Sec));
- }
- Address = FOTIterator->second;
- }
- uint32_t NumBlocks = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
-
+ uint32_t NumBlocksInBBRange = 0;
+ uint32_t NumBBRanges = 1;
+ typename ELFFile<ELFT>::uintX_t RangeBaseAddress = 0;
std::vector<BBAddrMap::BBEntry> BBEntries;
- uint32_t PrevBBEndOffset = 0;
- for (uint32_t BlockIndex = 0;
- !MetadataDecodeErr && !ULEBSizeErr && Cur && (BlockIndex < NumBlocks);
- ++BlockIndex) {
- uint32_t ID = Version >= 2
- ? readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr)
- : BlockIndex;
- uint32_t Offset = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
- uint32_t Size = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
- uint32_t MD = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
- if (Version >= 1) {
- // Offset is calculated relative to the end of the previous BB.
- Offset += PrevBBEndOffset;
- PrevBBEndOffset = Offset + Size;
+ if (FeatEnable.MultiBBRange) {
+ NumBBRanges = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ } else {
+ auto AddressOrErr = ExtractAddress();
+ if (!AddressOrErr)
+ return AddressOrErr.takeError();
+ RangeBaseAddress = *AddressOrErr;
+ NumBlocksInBBRange = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ }
+ std::vector<BBAddrMap::BBRangeEntry> BBRangeEntries;
+ uint32_t TotalNumBlocks = 0;
+ for (uint32_t BBRangeIndex = 0; BBRangeIndex < NumBBRanges;
+ ++BBRangeIndex) {
+ uint32_t PrevBBEndOffset = 0;
+ if (FeatEnable.MultiBBRange) {
+ auto AddressOrErr = ExtractAddress();
+ if (!AddressOrErr)
+ return AddressOrErr.takeError();
+ RangeBaseAddress = *AddressOrErr;
+ NumBlocksInBBRange = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
}
- Expected<BBAddrMap::BBEntry::Metadata> MetadataOrErr =
- BBAddrMap::BBEntry::Metadata::decode(MD);
- if (!MetadataOrErr) {
- MetadataDecodeErr = MetadataOrErr.takeError();
- break;
+ for (uint32_t BlockIndex = 0; !MetadataDecodeErr && !ULEBSizeErr && Cur &&
+ (BlockIndex < NumBlocksInBBRange);
+ ++BlockIndex) {
+ uint32_t ID = Version >= 2
+ ? readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr)
+ : BlockIndex;
+ uint32_t Offset = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ uint32_t Size = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ uint32_t MD = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ if (Version >= 1) {
+ // Offset is calculated relative to the end of the previous BB.
+ Offset += PrevBBEndOffset;
+ PrevBBEndOffset = Offset + Size;
+ }
+ Expected<BBAddrMap::BBEntry::Metadata> MetadataOrErr =
+ BBAddrMap::BBEntry::Metadata::decode(MD);
+ if (!MetadataOrErr) {
+ MetadataDecodeErr = MetadataOrErr.takeError();
+ break;
+ }
+ BBEntries.push_back({ID, Offset, Size, *MetadataOrErr});
}
- BBEntries.push_back({ID, Offset, Size, *MetadataOrErr});
+ TotalNumBlocks += BBEntries.size();
+ BBRangeEntries.push_back({RangeBaseAddress, std::move(BBEntries)});
}
- FunctionEntries.emplace_back(Address, std::move(BBEntries));
+ FunctionEntries.push_back({std::move(BBRangeEntries)});
- if (PGOAnalyses || FeatEnable.anyEnabled()) {
+ if (PGOAnalyses || FeatEnable.hasPGOAnalysis()) {
// Function entry count
uint64_t FuncEntryCount =
FeatEnable.FuncEntryCount
@@ -783,8 +819,8 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
std::vector<PGOAnalysisMap::PGOBBEntry> PGOBBEntries;
for (uint32_t BlockIndex = 0;
- (FeatEnable.BBFreq || FeatEnable.BrProb) && !MetadataDecodeErr &&
- !ULEBSizeErr && Cur && (BlockIndex < NumBlocks);
+ FeatEnable.hasPGOAnalysisBBData() && !MetadataDecodeErr &&
+ !ULEBSizeErr && Cur && (BlockIndex < TotalNumBlocks);
++BlockIndex) {
// Block frequency
uint64_t BBF = FeatEnable.BBFreq
diff --git a/llvm/lib/ObjectYAML/ELFEmitter.cpp b/llvm/lib/ObjectYAML/ELFEmitter.cpp
index b044d502c9975..3e32f8c89d841 100644
--- a/llvm/lib/ObjectYAML/ELFEmitter.cpp
+++ b/llvm/lib/ObjectYAML/ELFEmitter.cpp
@@ -1419,24 +1419,43 @@ void ELFState<ELFT>::writeSectionContent(
CBA.write(E.Feature);
SHeader.sh_size += 2;
}
-
- if (Section.PGOAnalyses) {
- if (E.Version < 2)
- WithColor::warning()
- << "unsupported SHT_LLVM_BB_ADDR_MAP version when using PGO: "
- << static_cast<int>(E.Version) << "; must use version >= 2";
+ auto FeatureOrErr = llvm::object::BBAddrMap::Features::decode(E.Feature);
+ bool MultiBBRangeFeatureEnabled = false;
+ if (!FeatureOrErr)
+ WithColor::warning() << toString(FeatureOrErr.takeError());
+ else
+ MultiBBRangeFeatureEnabled = FeatureOrErr->MultiBBRange;
+ bool MultiBBRange =
+ MultiBBRangeFeatureEnabled ||
+ (E.NumBBRanges.has_value() && E.NumBBRanges.value() > 1) ||
+ (E.BBRanges && E.BBRanges->size() > 1);
+ if (MultiBBRange && !MultiBBRangeFeatureEnabled)
+ WithColor::warning() << "feature value(" << E.Feature
+ << ") does not support multiple BB ranges.";
+ if (MultiBBRange) {
+ // Write the number of basic block ranges, which is overridden by the
+ // 'NumBBRanges' field when specified.
+ uint64_t NumBBRanges =
+ E.NumBBRanges.value_or(E.BBRanges ? E.BBRanges->size() : 0);
+ SHeader.sh_size += CBA.writeULEB128(NumBBRanges);
}
-
- // Write the address of the function.
- CBA.write<uintX_t>(E.Address, ELFT::TargetEndianness);
- // Write number of BBEntries (number of basic blocks in the function). This
- // is overridden by the 'NumBlocks' YAML field when specified.
- uint64_t NumBlocks =
- E.NumBlocks.value_or(E.BBEntries ? E.BBEntries->size() : 0);
- SHeader.sh_size += sizeof(uintX_t) + CBA.writeULEB128(NumBlocks);
- // Write all BBEntries.
- if (E.BBEntries) {
- for (const ELFYAML::BBAddrMapEntry::BBEntry &BBE : *E.BBEntries) {
+ if (!E.BBRanges)
+ continue;
+ uint64_t TotalNumBlocks = 0;
+ for (const ELFYAML::BBAddrMapEntry::BBRangeEntry &BBR : *E.BBRanges) {
+ // Write the base address of the range.
+ CBA.write<uintX_t>(BBR.BaseAddress, ELFT::TargetEndianness);
+ // Write number of BBEntries (number of basic blocks in this basic block
+ // range). This is overridden by the 'NumBlocks' YAML field when
+ // specified.
+ uint64_t NumBlocks =
+ BBR.NumBlocks.value_or(BBR.BBEntries ? BBR.BBEntries->size() : 0);
+ SHeader.sh_size += sizeof(uintX_t) + CBA.writeULEB128(NumBlocks);
+ // Write all BBEntries in this BBRange.
+ if (!BBR.BBEntries)
+ continue;
+ for (const ELFYAML::BBAddrMapEntry::BBEntry &BBE : *BBR.BBEntries) {
+ ++TotalNumBlocks;
if (Section.Type == llvm::ELF::SHT_LLVM_BB_ADDR_MAP && E.Version > 1)
SHeader.sh_size += CBA.writeULEB128(BBE.ID);
SHeader.sh_size += CBA.writeULEB128(BBE.AddressOffset);
@@ -1444,7 +1463,6 @@ void ELFState<ELFT>::writeSectionContent(
SHeader.sh_size += CBA.writeULEB128(BBE.Metadata);
}
}
-
if (!PGOAnalyses)
continue;
const ELFYAML::PGOAnalysisMapEntry &PGOEntry = PGOAnalyses->at(Idx);
@@ -1456,11 +1474,11 @@ void ELFState<ELFT>::writeSectionContent(
continue;
const auto &PGOBBEntries = PGOEntry.PGOBBEntries.value();
- if (!E.BBEntries || E.BBEntries->size() != PGOBBEntries.size()) {
+ if (TotalNumBlocks != PGOBBEntries.size()) {
WithColor::warning() << "PBOBBEntries must be the same length as "
"BBEntries in SHT_LLVM_BB_ADDR_MAP.\n"
<< "Mismatch on function with address: "
- << E.Address;
+ << E.getFunctionAddress();
continue;
}
diff --git a/llvm/lib/ObjectYAML/ELFYAML.cpp b/llvm/lib/ObjectYAML/ELFYAML.cpp
index 6ad4a067415ac..31e90fea6e46a 100644
--- a/llvm/lib/ObjectYAML/ELFYAML.cpp
+++ b/llvm/lib/ObjectYAML/ELFYAML.cpp
@@ -1680,7 +1680,6 @@ void MappingTraits<std::unique_ptr<ELFYAML::Chunk>>::mapping(
Section.reset(new ELFYAML::CallGraphProfileSection());
sectionMapping(IO, *cast<ELFYAML::CallGraphProfileSection>(Section.get()));
break;
- case ELF::SHT_LLVM_BB_ADDR_MAP_V0:
case ELF::SHT_LLVM_BB_ADDR_MAP:
if (!IO.outputting())
Section.reset(new ELFYAML::BBAddrMapSection());
@@ -1812,7 +1811,13 @@ void MappingTraits<ELFYAML::BBAddrMapEntry>::mapping(
assert(IO.getContext() && "The IO context is not initialized");
IO.mapRequired("Version", E.Version);
IO.mapOptional("Feature", E.Feature, Hex8(0));
- IO.mapOptional("Address", E.Address, Hex64(0));
+ IO.mapOptional("NumBBRanges", E.NumBBRanges);
+ IO.mapOptional("BBRanges", E.BBRanges);
+}
+
+void MappingTraits<ELFYAML::BBAddrMapEntry::BBRangeEntry>::mapping(
+ IO &IO, ELFYAML::BBAddrMapEntry::BBRangeEntry &E) {
+ IO.mapOptional("BaseAddress", E.BaseAddress, Hex64(0));
IO.mapOptional("NumBlocks", E.NumBlocks);
IO.mapOptional("BBEntries", E.BBEntries);
}
diff --git a/llvm/test/CodeGen/X86/basic-block-sections-labels-functions-sections.ll b/llvm/test/CodeGen/X86/basic-block-address-map-function-sections.ll
similarity index 94%
rename from llvm/test/CodeGen/X86/basic-block-sections-labels-functions-sections.ll
rename to llvm/test/CodeGen/X86/basic-block-address-map-function-sections.ll
index 3be3ab70f8a69..d7678604cffa2 100644
--- a/llvm/test/CodeGen/X86/basic-block-sections-labels-functions-sections.ll
+++ b/llvm/test/CodeGen/X86/basic-block-address-map-function-sections.ll
@@ -1,4 +1,5 @@
; RUN: llc < %s -mtriple=x86_64 -function-sections -basic-block-sections=labels | FileCheck %s
+; RUN: llc < %s -mtriple=x86_64 -function-sections -basic-block-address-map | FileCheck %s
$_Z4fooTIiET_v = comdat any
diff --git a/llvm/test/CodeGen/X86/basic-block-address-map-with-basic-block-sections.ll b/llvm/test/CodeGen/X86/basic-block-address-map-with-basic-block-sections.ll
new file mode 100644
index 0000000000000..6354e2ec6889f
--- /dev/null
+++ b/llvm/test/CodeGen/X86/basic-block-address-map-with-basic-block-sections.ll
@@ -0,0 +1,71 @@
+;; Check that the -basic-block-address-map option works when used along with -basic-block-sections.
+;; Let a function with 4 basic blocks get split into 2 sections.
+; RUN: echo '!_Z3bazb' > %t
+; RUN: echo '!!0 2' >> %t
+; RUN: llc < %s -mtriple=x86_64 -basic-block-address-map -basic-block-sections=%t | FileCheck %s
+
+define void @_Z3bazb(i1 zeroext) personality i32 (...)* @__gxx_personality_v0 {
+ br i1 %0, label %2, label %7
+
+2:
+ %3 = invoke i32 @_Z3barv()
+ to label %7 unwind label %5
+ br label %9
+
+5:
+ landingpad { i8*, i32 }
+ catch i8* null
+ br label %9
+
+7:
+ %8 = call i32 @_Z3foov()
+ br label %9
+
+9:
+ ret void
+}
+
+declare i32 @_Z3barv() #1
+
+declare i32 @_Z3foov() #1
+
+declare i32 @__gxx_personality_v0(...)
+
+; CHECK: .text
+; CHECK-LABEL: _Z3bazb:
+; CHECK-LABEL: .Lfunc_begin0:
+; CHECK-LABEL: .LBB_END0_0:
+; CHECK-LABEL: .LBB0_1:
+; CHECK-LABEL: .LBB_END0_1:
+; CHECK: .section .text.split._Z3bazb,"ax", at progbits
+; CHECK-LABEL: _Z3bazb.cold:
+; CHECK-LABEL: .LBB_END0_2:
+; CHECK-LABEL: .LBB0_3:
+; CHECK-LABEL: .LBB_END0_3:
+; CHECK-LABEL: .Lfunc_end0:
+
+; CHECK: .section .llvm_bb_addr_map,"o", at llvm_bb_addr_map,.text.hot._Z3bazb
+; CHECK-NEXT: .byte 2 # version
+; CHECK-NEXT: .byte 8 # feature
+; CHECK-NEXT: .byte 2 # number of basic block ranges
+; CHECK-NEXT: .quad .Lfunc_begin0 # base address
+; CHECK-NEXT: .byte 2 # number of basic blocks
+; CHECK-NEXT: .byte 0 # BB id
+; CHECK-NEXT: .uleb128 .Lfunc_begin0-.Lfunc_begin0
+; CHECK-NEXT: .uleb128 .LBB_END0_0-.Lfunc_begin0
+; CHECK-NEXT: .byte 0
+; CHECK-NEXT: .byte 2 # BB id
+; CHECK-NEXT: .uleb128 .LBB0_1-.LBB_END0_0
+; CHECK-NEXT: .uleb128 .LBB_END0_1-.LBB0_1
+; CHECK-NEXT: .byte 5
+; CHECK-NEXT: .quad _Z3bazb.cold # base address
+; CHECK-NEXT: .byte 2 # number of basic blocks
+; CHECK-NEXT: .byte 1 # BB id
+; CHECK-NEXT: .uleb128 _Z3bazb.cold-_Z3bazb.cold
+; CHECK-NEXT: .uleb128 .LBB_END0_2-_Z3bazb.cold
+; CHECK-NEXT: .byte 8
+; CHECK-NEXT: .byte 3 # BB id
+; CHECK-NEXT: .uleb128 .LBB0_3-.LBB_END0_2
+; CHECK-NEXT: .uleb128 .LBB_END0_3-.LBB0_3
+; CHECK-NEXT: .byte 1
+
diff --git a/llvm/test/CodeGen/X86/basic-block-address-map-with-mfs.ll b/llvm/test/CodeGen/X86/basic-block-address-map-with-mfs.ll
new file mode 100644
index 0000000000000..2565db23c0249
--- /dev/null
+++ b/llvm/test/CodeGen/X86/basic-block-address-map-with-mfs.ll
@@ -0,0 +1,89 @@
+; COM: Emitting basic-block-address-map when machine function splitting is enabled.
+; RUN: llc < %s -mtriple=x86_64 -function-sections -split-machine-functions -basic-block-address-map | FileCheck %s --check-prefixes=CHECK,BASIC
+
+; COM: Emitting basic-block-address-map with PGO analysis with machine function splitting enabled.
+; RUN: llc < %s -mtriple=x86_64 -function-sections -split-machine-functions -basic-block-address-map -pgo-analysis-map=func-entry-count,bb-freq,br-prob | FileCheck %s --check-prefixes=CHECK,PGO
+
+define void @foo(i1 zeroext %0) nounwind !prof !14 {
+ br i1 %0, label %2, label %4, !prof !15
+
+2: ; preds = %1
+ %3 = call i32 @bar()
+ br label %6
+
+4: ; preds = %1
+ %5 = call i32 @baz()
+ br label %6
+
+6: ; preds = %4, %2
+ %7 = tail call i32 @qux()
+ ret void
+}
+
+declare i32 @bar()
+declare i32 @baz()
+declare i32 @qux()
+
+!llvm.module.flags = !{!0}
+!0 = !{i32 1, !"ProfileSummary", !1}
+!1 = !{!2, !3, !4, !5, !6, !7, !8, !9}
+!2 = !{!"ProfileFormat", !"InstrProf"}
+!3 = !{!"TotalCount", i64 10000}
+!4 = !{!"MaxCount", i64 10}
+!5 = !{!"MaxInternalCount", i64 1}
+!6 = !{!"MaxFunctionCount", i64 1000}
+!7 = !{!"NumCounts", i64 3}
+!8 = !{!"NumFunctions", i64 5}
+!9 = !{!"DetailedSummary", !10}
+!10 = !{!11, !12, !13}
+!11 = !{i32 10000, i64 100, i32 1}
+!12 = !{i32 999900, i64 100, i32 1}
+!13 = !{i32 999999, i64 1, i32 2}
+!14 = !{!"function_entry_count", i64 7000}
+!15 = !{!"branch_weights", i32 7000, i32 0}
+
+; CHECK: .section .text.hot.foo,"ax", at progbits
+; CHECK-LABEL: foo:
+; CHECK-LABEL: .Lfunc_begin0:
+; CHECK-LABEL: .LBB_END0_0:
+; CHECK-LABEL: .LBB0_1:
+; CHECK-LABEL: .LBB_END0_1:
+; CHECK: .section .text.split.foo,"ax", at progbits
+; CHECK-LABEL: foo.cold:
+; CHECK-LABEL: .LBB_END0_2:
+; CHECK-LABEL: .Lfunc_end0:
+
+; CHECK: .section .llvm_bb_addr_map,"o", at llvm_bb_addr_map,.text.hot.foo
+; CHECK-NEXT: .byte 2 # version
+; BASIC-NEXT: .byte 8 # feature
+; PGO-NEXT: .byte 15 # feature
+; CHECK-NEXT: .byte 2 # number of basic block ranges
+; CHECK-NEXT: .quad .Lfunc_begin0 # base address
+; CHECK-NEXT: .byte 2 # number of basic blocks
+; CHECK-NEXT: .byte 0 # BB id
+; CHECK-NEXT: .uleb128 .Lfunc_begin0-.Lfunc_begin0
+; CHECK-NEXT: .uleb128 .LBB_END0_0-.Lfunc_begin0
+; CHECK-NEXT: .byte 8
+; CHECK-NEXT: .byte 1 # BB id
+; CHECK-NEXT: .uleb128 .LBB0_1-.LBB_END0_0
+; CHECK-NEXT: .uleb128 .LBB_END0_1-.LBB0_1
+; CHECK-NEXT: .byte 3
+; CHECK-NEXT: .quad foo.cold # base address
+; CHECK-NEXT: .byte 1 # number of basic blocks
+; CHECK-NEXT: .byte 2 # BB id
+; CHECK-NEXT: .uleb128 foo.cold-foo.cold
+; CHECK-NEXT: .uleb128 .LBB_END0_2-foo.cold
+; CHECK-NEXT: .byte 3
+
+;; PGO Analysis Map
+; PGO: .ascii "\3306" # function entry count
+; PGO-NEXT: .ascii "\200\200\200\200\200\200\200 " # basic block frequency
+; PGO-NEXT: .byte 2 # basic block successor count
+; PGO-NEXT: .byte 1 # successor BB ID
+; PGO-NEXT: .ascii "\200\200\200\200\b" # successor branch probability
+; PGO-NEXT: .byte 2 # successor BB ID
+; PGO-NEXT: .byte 0 # successor branch probability
+; PGO-NEXT: .ascii "\200\200\200\374\377\377\377\037" # basic block frequency
+; PGO-NEXT: .byte 0 # basic block successor count
+; PGO-NEXT: .ascii "\200\200\200\004" # basic block frequency
+; PGO-NEXT: .byte 0 # basic block successor count
diff --git a/llvm/test/CodeGen/X86/basic-block-sections-labels.ll b/llvm/test/CodeGen/X86/basic-block-address-map.ll
similarity index 90%
rename from llvm/test/CodeGen/X86/basic-block-sections-labels.ll
rename to llvm/test/CodeGen/X86/basic-block-address-map.ll
index 0b0b00a0b9814..6ab24b494936a 100644
--- a/llvm/test/CodeGen/X86/basic-block-sections-labels.ll
+++ b/llvm/test/CodeGen/X86/basic-block-address-map.ll
@@ -1,5 +1,7 @@
; Check the basic block sections labels option
+; RUN: llc < %s -mtriple=x86_64 -function-sections -unique-section-names=true -basic-block-address-map | FileCheck %s --check-prefixes=CHECK,UNIQ
; RUN: llc < %s -mtriple=x86_64 -function-sections -unique-section-names=true -basic-block-sections=labels | FileCheck %s --check-prefixes=CHECK,UNIQ
+; RUN: llc < %s -mtriple=x86_64 -function-sections -unique-section-names=false -basic-block-address-map | FileCheck %s --check-prefixes=CHECK,NOUNIQ
; RUN: llc < %s -mtriple=x86_64 -function-sections -unique-section-names=false -basic-block-sections=labels | FileCheck %s --check-prefixes=CHECK,NOUNIQ
; RUN: llc < %s -mtriple=x86_64 -function-sections -unique-section-names=true -basic-block-sections=labels -split-machine-functions | FileCheck %s --check-prefixes=CHECK,UNIQ
diff --git a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
index 6864932411aec..cc977af76dfb4 100644
--- a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
+++ b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
@@ -90,40 +90,42 @@ Sections:
Link: .text.foo
Entries:
- Version: 2
- Address: [[FOO_ADDR]]
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - ID: 1
- AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x1
- Size: 0x4
- Metadata: 0x0
- - ID: 5
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: [[FOO_ADDR]]
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 5
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
- Name: .llvm_bb_addr_map.bar
Type: SHT_LLVM_BB_ADDR_MAP
Link: .text.bar
Entries:
- Version: 1
- Address: [[BAR_ADDR]]
- BBEntries:
- - AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - AddressOffset: 0x4
- Size: 0x2
- Metadata: 0x0
- - AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
+ BBRanges:
+ - BaseAddress: [[BAR_ADDR]]
+ BBEntries:
+ - AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - AddressOffset: 0x4
+ Size: 0x2
+ Metadata: 0x0
+ - AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
Symbols:
- Name: foo
@@ -164,36 +166,38 @@ Sections:
Link: .text.foo
Entries:
- Version: 2
- Address: 0x4000
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - ID: 1
- AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x1
- Size: 0x4
- Metadata: 0x0
- - ID: 5
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: 0x4000
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 5
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
- Version: 1
- Address: 0x5000
- BBEntries:
- - AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - AddressOffset: 0x4
- Size: 0x2
- Metadata: 0x0
- - AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
+ BBRanges:
+ - BaseAddress: 0x5000
+ BBEntries:
+ - AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - AddressOffset: 0x4
+ Size: 0x2
+ Metadata: 0x0
+ - AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
Symbols:
- Name: foo
diff --git a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
index f796963481cd4..88afaceac8c51 100644
--- a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
+++ b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
@@ -24,17 +24,19 @@ Sections:
Link: .text
Entries:
- Version: 2
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
+ BBRanges:
+ - BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
- Version: 2
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
+ BBRanges:
+ - BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
- Name: .rela.llvm_bb_addr_map
Type: SHT_RELA
Flags: [ SHF_INFO_LINK ]
diff --git a/llvm/test/tools/llvm-objdump/X86/elf-pgoanalysismap.yaml b/llvm/test/tools/llvm-objdump/X86/elf-pgoanalysismap.yaml
index c4bf443f920a2..732fab3e2a378 100644
--- a/llvm/test/tools/llvm-objdump/X86/elf-pgoanalysismap.yaml
+++ b/llvm/test/tools/llvm-objdump/X86/elf-pgoanalysismap.yaml
@@ -25,13 +25,14 @@ Sections:
Link: .text.foo
Entries:
- Version: 2
- Address: 0x0
Feature: 0x1
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
+ BBRanges:
+ - BaseAddress: 0x0
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
PGOAnalyses:
- FuncEntryCount: 1000
Symbols:
@@ -65,25 +66,26 @@ Sections:
Link: .text.foo
Entries:
- Version: 2
- Address: 0x0
Feature: 0x3
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - ID: 1
- AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x1
- Size: 0x4
- Metadata: 0x0
- - ID: 5
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: 0x0
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 5
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
PGOAnalyses:
- FuncEntryCount: 1000
PGOBBEntries:
@@ -126,25 +128,26 @@ Sections:
Link: .text.foo
Entries:
- Version: 2
- Address: 0x0
Feature: 0x7
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - ID: 1
- AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x1
- Size: 0x4
- Metadata: 0x0
- - ID: 5
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: 0x0
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 5
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
PGOAnalyses:
- FuncEntryCount: 1000
PGOBBEntries:
diff --git a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-relocatable.test b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-relocatable.test
index e6b6cc344a8e7..e7f78491a9473 100644
--- a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-relocatable.test
+++ b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-relocatable.test
@@ -8,32 +8,42 @@
# CHECK-NEXT: Function {
# CHECK-NEXT: At: 0x0
# CHECK-NEXT: Name: <?>
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 0
-# CHECK-NEXT: Offset: 0x0
-# CHECK-NEXT: Size: 0xF
-# CHECK-NEXT: HasReturn: Yes
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: No
-# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: Base Address: 0x0
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 0
+# CHECK-NEXT: Offset: 0x0
+# CHECK-NEXT: Size: 0xF
+# CHECK-NEXT: HasReturn: Yes
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: Function {
# CHECK-NEXT: At: 0x10
# CHECK-NEXT: Name: <?>
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 0
-# CHECK-NEXT: Offset: 0x0
-# CHECK-NEXT: Size: 0x11
-# CHECK-NEXT: HasReturn: No
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: Yes
-# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: Base Address: 0x10
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 0
+# CHECK-NEXT: Offset: 0x0
+# CHECK-NEXT: Size: 0x11
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: Yes
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: }
@@ -54,17 +64,19 @@ Sections:
Link: .text
Entries:
- Version: 2
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0xF
- Metadata: 0x1
+ BBRanges:
+ - BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0xF
+ Metadata: 0x1
- Version: 2
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x11
- Metadata: 0x8
+ BBRanges:
+ - BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x11
+ Metadata: 0x8
- Name: .rela.llvm_bb_addr_map
Type: SHT_RELA
Flags: [ SHF_INFO_LINK ]
@@ -128,7 +140,8 @@ Sections:
Link: .text
Entries:
- Version: 2
- BBEntries:
+ BBRanges:
+ - BBEntries:
- ID: 0
AddressOffset: 0x0
Size: 0xF
@@ -181,12 +194,13 @@ Sections:
Flags: [ SHF_LINK_ORDER ]
Entries:
- Version: 2
- Address: 0xF
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0xF
- Metadata: 0x1
+ BBRanges:
+ - BaseAddress: 0xF
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0xF
+ Metadata: 0x1
# RUN: yaml2obj %s --docnum=5 -o %t5.o
# RUN: llvm-readobj %t5.o --bb-addr-map 2>&1 | FileCheck %s --check-prefix=ET-DYN-NO-WARNING -DFILE=%t5.o
@@ -195,16 +209,21 @@ Sections:
# ET-DYN-NO-WARNING: Function {
# ET-DYN-NO-WARNING: At: 0xF
# ET-DYN-NO-WARNING: Name: <?>
-# ET-DYN-NO-WARNING: BB entries [
+# ET-DYN-NO-WARNING: BB Ranges [
# ET-DYN-NO-WARNING: {
-# ET-DYN-NO-WARNING: ID: 0
-# ET-DYN-NO-WARNING: Offset: 0x0
-# ET-DYN-NO-WARNING: Size: 0xF
-# ET-DYN-NO-WARNING: HasReturn: Yes
-# ET-DYN-NO-WARNING: HasTailCall: No
-# ET-DYN-NO-WARNING: IsEHPad: No
-# ET-DYN-NO-WARNING: CanFallThrough: No
-# ET-DYN-NO-WARNING: HasIndirectBranch: No
+# ET-DYN-NO-WARNING: Base Address: 0xF
+# ET-DYN-NO-WARNING: BB Entries [
+# ET-DYN-NO-WARNING: {
+# ET-DYN-NO-WARNING: ID: 0
+# ET-DYN-NO-WARNING: Offset: 0x0
+# ET-DYN-NO-WARNING: Size: 0xF
+# ET-DYN-NO-WARNING: HasReturn: Yes
+# ET-DYN-NO-WARNING: HasTailCall: No
+# ET-DYN-NO-WARNING: IsEHPad: No
+# ET-DYN-NO-WARNING: CanFallThrough: No
+# ET-DYN-NO-WARNING: HasIndirectBranch: No
+# ET-DYN-NO-WARNING: }
+# ET-DYN-NO-WARNING: ]
# ET-DYN-NO-WARNING: }
# ET-DYN-NO-WARNING: ]
# ET-DYN-NO-WARNING: }
diff --git a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
index 0593f04d6e301..c5d071c11d1de 100644
--- a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
+++ b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map.test
@@ -15,8 +15,8 @@
# RUN: llvm-readelf %t1.x32.o --bb-addr-map | FileCheck %s --check-prefix=GNU
## Check that a malformed section can be handled.
-# RUN: yaml2obj --docnum=1 %s -DBITS=32 -DSIZE=6 -o %t2.o
-# RUN: llvm-readobj %t2.o --bb-addr-map 2>&1 | FileCheck %s -DOFFSET=0x00000006 -DFILE=%t2.o --check-prefix=TRUNCATED
+# RUN: yaml2obj --docnum=1 %s -DBITS=32 -DSIZE=7 -o %t2.o
+# RUN: llvm-readobj %t2.o --bb-addr-map 2>&1 | FileCheck %s -DOFFSET=0x00000007 -DFILE=%t2.o --check-prefix=TRUNCATED
## Check that invalid metadata can be handled.
# RUN: yaml2obj --docnum=1 %s -DBITS=32 -DMETADATA=0xF000002 -o %t3.o
@@ -27,42 +27,57 @@
# CHECK-NEXT: At: [[ADDR]]
# CHECK-NEXT: warning: '[[FILE]]': could not identify function symbol for address ([[ADDR]]) in SHT_LLVM_BB_ADDR_MAP section with index 3
# CHECK-NEXT: Name: <?>
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 0
-# CHECK-NEXT: Offset: 0x0
-# CHECK-NEXT: Size: 0x1
-# CHECK-NEXT: HasReturn: No
-# CHECK-NEXT: HasTailCall: Yes
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: No
-# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: Base Address: [[ADDR]]
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 0
+# CHECK-NEXT: Offset: 0x0
+# CHECK-NEXT: Size: 0x1
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: Yes
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 2
-# CHECK-NEXT: Offset: 0x4
-# CHECK-NEXT: Size: 0x4
-# CHECK-NEXT: HasReturn: Yes
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: Yes
-# CHECK-NEXT: CanFallThrough: No
-# CHECK-NEXT: HasIndirectBranch: Yes
+# CHECK-NEXT: Base Address: 0x44444
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 2
+# CHECK-NEXT: Offset: 0x3
+# CHECK-NEXT: Size: 0x4
+# CHECK-NEXT: HasReturn: Yes
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: Yes
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: Yes
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: Function {
# CHECK-NEXT: At: 0x22222
# CHECK-NEXT: Name: foo
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 4
-# CHECK-NEXT: Offset: 0x6
-# CHECK-NEXT: Size: 0x7
-# CHECK-NEXT: HasReturn: No
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: Yes
-# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: Base Address: 0x22222
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 4
+# CHECK-NEXT: Offset: 0x6
+# CHECK-NEXT: Size: 0x7
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: Yes
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: }
@@ -78,26 +93,31 @@
# TRUNCATED-NEXT: Function {
# TRUNCATED-NEXT: At: 0x33333
# TRUNCATED-NEXT: Name: bar
-# TRUNCATED-NEXT: BB entries [
+# TRUNCATED-NEXT: BB Ranges [
# TRUNCATED-NEXT: {
-# TRUNCATED-NEXT: ID: 6
-# TRUNCATED-NEXT: Offset: 0x9
-# TRUNCATED-NEXT: Size: 0xA
-# TRUNCATED-NEXT: HasReturn: Yes
-# TRUNCATED-NEXT: HasTailCall: Yes
-# TRUNCATED-NEXT: IsEHPad: No
-# TRUNCATED-NEXT: CanFallThrough: Yes
-# TRUNCATED-NEXT: HasIndirectBranch: Yes
-# TRUNCATED-NEXT: }
-# TRUNCATED-NEXT: {
-# TRUNCATED-NEXT: ID: 7
-# TRUNCATED-NEXT: Offset: 0x1F
-# TRUNCATED-NEXT: Size: 0xD
-# TRUNCATED-NEXT: HasReturn: No
-# TRUNCATED-NEXT: HasTailCall: Yes
-# TRUNCATED-NEXT: IsEHPad: Yes
-# TRUNCATED-NEXT: CanFallThrough: Yes
-# TRUNCATED-NEXT: HasIndirectBranch: No
+# TRUNCATED-NEXT: Base Address: 0x33333
+# TRUNCATED-NEXT: BB Entries [
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 6
+# TRUNCATED-NEXT: Offset: 0x9
+# TRUNCATED-NEXT: Size: 0xA
+# TRUNCATED-NEXT: HasReturn: Yes
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: No
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: Yes
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 7
+# TRUNCATED-NEXT: Offset: 0x1F
+# TRUNCATED-NEXT: Size: 0xD
+# TRUNCATED-NEXT: HasReturn: No
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: Yes
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: No
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: ]
# TRUNCATED-NEXT: }
# TRUNCATED-NEXT: ]
# TRUNCATED-NEXT: }
@@ -124,23 +144,28 @@ Sections:
Link: .text
Entries:
- Version: 2
- Address: [[ADDR=0x11111]]
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x1
- Metadata: [[METADATA=0x2]]
- - ID: 2
- AddressOffset: 0x3
- Size: 0x4
- Metadata: 0x15
+ Feature: 0x8
+ BBRanges:
+ - BaseAddress: [[ADDR=0x11111]]
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: [[METADATA=0x2]]
+ - BaseAddress: 0x44444
+ BBEntries:
+ - ID: 2
+ AddressOffset: 0x3
+ Size: 0x4
+ Metadata: 0x15
- Version: 2
- Address: 0x22222
- BBEntries:
- - ID: 4
- AddressOffset: 0x6
- Size: 0x7
- Metadata: 0x8
+ BBRanges:
+ - BaseAddress: 0x22222
+ BBEntries:
+ - ID: 4
+ AddressOffset: 0x6
+ Size: 0x7
+ Metadata: 0x8
- Name: dummy_section
Type: SHT_PROGBITS
Size: 16
@@ -149,16 +174,17 @@ Sections:
Link: .text.bar
Entries:
- Version: 2
- Address: 0x33333
- BBEntries:
- - ID: 6
- AddressOffset: 0x9
- Size: 0xa
- Metadata: 0x1b
- - ID: 7
- AddressOffset: 0xc
- Size: 0xd
- Metadata: 0xe
+ BBRanges:
+ - BaseAddress: 0x33333
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0x9
+ Size: 0xa
+ Metadata: 0x1b
+ - ID: 7
+ AddressOffset: 0xc
+ Size: 0xd
+ Metadata: 0xe
Symbols:
- Name: foo
Section: .text
@@ -168,103 +194,3 @@ Symbols:
Section: .text.bar
Type: STT_FUNC
Value: 0x33333
-
-## Check that using the SHT_LLVM_BB_ADDR_MAP_V0 section type generates the same
-## result as using the SHT_LLVM_BB_ADDR_MAP section type with Version=0.
-## The Version field is required even for SHT_LLVM_BB_ADDR_MAP_V0 but it
-## should not impact the result. This unideal behavior will be gone once
-## SHT_LLVM_BB_ADDR_MAP_V0 is deprecated.
-
-# RUN: yaml2obj --docnum=2 %s -DVERSION=255 -DSECTION_TYPE=SHT_LLVM_BB_ADDR_MAP_V0 -o %t2.type0
-# RUN: llvm-readobj %t2.type0 --bb-addr-map 2>&1 | FileCheck %s --check-prefix=V0
-
-# RUN: yaml2obj --docnum=2 %s -DVERSION=0 -DSECTION_TYPE=SHT_LLVM_BB_ADDR_MAP -o %t2.version0
-# RUN: llvm-readobj %t2.version0 --bb-addr-map 2>&1 | FileCheck %s --check-prefix=V0
-
-# V0: BBAddrMap [
-# V0-NEXT: Function {
-# V0-NEXT: At: 0x11111
-# V0-NEXT: Name: foo
-# V0-NEXT: BB entries [
-# V0-NEXT: {
-# V0-NEXT: ID: 0
-# V0-NEXT: Offset: 0x1
-# V0-NEXT: Size: 0x2
-# V0-NEXT: HasReturn:
-# V0-NEXT: HasTailCall:
-# V0-NEXT: IsEHPad:
-# V0-NEXT: CanFallThrough:
-# V0-NEXT: HasIndirectBranch:
-# V0-NEXT: }
-# V0-NEXT: {
-# V0-NEXT: ID: 1
-# V0-NEXT: Offset: 0x4
-# V0-NEXT: Size: 0x5
-# V0-NEXT: HasReturn:
-# V0-NEXT: HasTailCall:
-# V0-NEXT: IsEHPad:
-# V0-NEXT: CanFallThrough:
-# V0-NEXT: HasIndirectBranch:
-# V0-NEXT: }
-# V0-NEXT: ]
-# V0-NEXT: }
-
-## Check version 1 (without BB IDs).
-# RUN: yaml2obj --docnum=2 %s -DVERSION=1 -DSECTION_TYPE=SHT_LLVM_BB_ADDR_MAP -o %t3
-# RUN: llvm-readobj %t3 --bb-addr-map 2>&1 | FileCheck %s --check-prefix=V1
-
-# V1: BBAddrMap [
-# V1-NEXT: Function {
-# V1-NEXT: At: 0x11111
-# V1-NEXT: Name: foo
-# V1-NEXT: BB entries [
-# V1-NEXT: {
-# V1-NEXT: ID: 0
-# V1-NEXT: Offset: 0x1
-# V1-NEXT: Size: 0x2
-# V1-NEXT: HasReturn:
-# V1-NEXT: HasTailCall:
-# V1-NEXT: IsEHPad:
-# V1-NEXT: CanFallThrough:
-# V1-NEXT: HasIndirectBranch:
-# V1-NEXT: }
-# V1-NEXT: {
-# V1-NEXT: ID: 1
-# V1-NEXT: Offset: 0x7
-# V1-NEXT: Size: 0x5
-# V1-NEXT: HasReturn:
-# V1-NEXT: HasTailCall:
-# V1-NEXT: IsEHPad:
-# V1-NEXT: CanFallThrough:
-# V1-NEXT: HasIndirectBranch:
-# V1-NEXT: }
-# V1-NEXT: ]
-# V1-NEXT: }
-
---- !ELF
-FileHeader:
- Class: ELFCLASS64
- Data: ELFDATA2LSB
- Type: ET_EXEC
-Sections:
- - Name: .text.foo
- Type: SHT_PROGBITS
- Flags: [SHF_ALLOC]
- - Name: .llvm_bb_addr_map
- Type: [[SECTION_TYPE]]
- Link: .text.foo
- Entries:
- - Version: [[VERSION]]
- Address: 0x11111
- BBEntries:
- - AddressOffset: 0x1
- Size: 0x2
- Metadata: 0x3
- - AddressOffset: 0x4
- Size: 0x5
- Metadata: 0x6
-Symbols:
- - Name: foo
- Section: .text.foo
- Type: STT_FUNC
- Value: 0x11111
diff --git a/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml b/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
index 629c29e202aeb..f9b3c30c158e4 100644
--- a/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
+++ b/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
@@ -14,30 +14,31 @@
# VALID-NEXT: - Name: .llvm_bb_addr_map
# VALID-NEXT: Type: SHT_LLVM_BB_ADDR_MAP
# VALID-NEXT: Entries:
-## The 'Address' field is omitted when it's zero.
# VALID-NEXT: - Version: 2
-# VALID-NEXT: Feature: 0xFF
-# VALID-NEXT: BBEntries:
-# VALID-NEXT: - ID: 0
-# VALID-NEXT: AddressOffset: 0x1
-# VALID-NEXT: Size: 0x2
-# VALID-NEXT: Metadata: 0x3
-# VALID-NEXT: - ID: 2
-# VALID-NEXT: AddressOffset: 0x4
-# VALID-NEXT: Size: 0x5
-# VALID-NEXT: Metadata: 0x6
-# VALID-NEXT: - ID: 4
-# VALID-NEXT: AddressOffset: 0xFFFFFFFFFFFFFFF7
-# VALID-NEXT: Size: 0xFFFFFFFFFFFFFFF8
-# VALID-NEXT: Metadata: 0xFFFFFFFFFFFFFFF9
+# VALID-NEXT: BBRanges:
+## The 'BaseAddress' field is omitted when it's zero.
+# VALID-NEXT: - BBEntries:
+# VALID-NEXT: - ID: 0
+# VALID-NEXT: AddressOffset: 0x1
+# VALID-NEXT: Size: 0x2
+# VALID-NEXT: Metadata: 0x3
+# VALID-NEXT: - ID: 2
+# VALID-NEXT: AddressOffset: 0x4
+# VALID-NEXT: Size: 0x5
+# VALID-NEXT: Metadata: 0x6
+# VALID-NEXT: - ID: 4
+# VALID-NEXT: AddressOffset: 0xFFFFFFFFFFFFFFF7
+# VALID-NEXT: Size: 0xFFFFFFFFFFFFFFF8
+# VALID-NEXT: Metadata: 0xFFFFFFFFFFFFFFF9
# VALID-NEXT: - Version: 2
-# VALID-NEXT: Feature: 0xEE
-# VALID-NEXT: Address: 0xFFFFFFFFFFFFFF20
-# VALID-NEXT: BBEntries:
-# VALID-NEXT: - ID: 6
-# VALID-NEXT: AddressOffset: 0xA
-# VALID-NEXT: Size: 0xB
-# VALID-NEXT: Metadata: 0xC
+# VALID-NEXT: Feature: 0x8
+# VALID-NEXT: BBRanges:
+# VALID-NEXT: - BaseAddress: 0xFFFFFFFFFFFFFF20
+# VALID-NEXT: BBEntries:
+# VALID-NEXT: - ID: 6
+# VALID-NEXT: AddressOffset: 0xA
+# VALID-NEXT: Size: 0xB
+# VALID-NEXT: Metadata: 0xC
--- !ELF
FileHeader:
@@ -50,30 +51,33 @@ Sections:
ShSize: [[SIZE=<none>]]
Entries:
- Version: 2
- Feature: 0xFF
- Address: 0x0
- BBEntries:
- - ID: 0
- AddressOffset: 0x1
- Size: 0x2
- Metadata: 0x3
- - ID: 2
- AddressOffset: 0x4
- Size: 0x5
- Metadata: 0x6
- - ID: 4
- AddressOffset: 0xFFFFFFFFFFFFFFF7
- Size: 0xFFFFFFFFFFFFFFF8
- Metadata: 0xFFFFFFFFFFFFFFF9
+ Feature: 0x0
+ BBRanges:
+ - BaseAddress: 0x0
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x1
+ Size: 0x2
+ Metadata: 0x3
+ - ID: 2
+ AddressOffset: 0x4
+ Size: 0x5
+ Metadata: 0x6
+ - ID: 4
+ AddressOffset: 0xFFFFFFFFFFFFFFF7
+ Size: 0xFFFFFFFFFFFFFFF8
+ Metadata: 0xFFFFFFFFFFFFFFF9
- Version: 2
- Feature: 0xEE
- Address: 0xFFFFFFFFFFFFFF20
- NumBlocks: [[NUMBLOCKS=<none>]]
- BBEntries:
- - ID: 6
- AddressOffset: 0xA
- Size: 0xB
- Metadata: 0xC
+ Feature: 0x8
+ NumBBRanges: [[NUMBBRANGES=<none>]]
+ BBRanges:
+ - BaseAddress: 0xFFFFFFFFFFFFFF20
+ NumBlocks: [[NUMBLOCKS=<none>]]
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0xA
+ Size: 0xB
+ Metadata: 0xC
## Check obj2yaml can dump empty .llvm_bb_addr_map sections.
@@ -114,19 +118,21 @@ Sections:
# MULTI-NEXT: - Name: .llvm_bb_addr_map
# MULTI-NEXT: Type: SHT_LLVM_BB_ADDR_MAP
# MULTI-NEXT: Entries:
-## Fields 'Address' and 'Feature' are omitted when they are zero.
+## Fields 'BaseAddress' and 'Feature' are omitted when they are zero.
# MULTI-NEXT: - Version: 0
-# MULTI-NEXT: BBEntries:
-# MULTI-NEXT: - ID: 0
-# MULTI-NEXT: AddressOffset: 0x1
-# MULTI-NEXT: Size: 0x2
-# MULTI-NEXT: Metadata: 0x3
+# MULTI-NEXT: BBRanges:
+# MULTI-NEXT: - BBEntries:
+# MULTI-NEXT: - ID: 0
+# MULTI-NEXT: AddressOffset: 0x1
+# MULTI-NEXT: Size: 0x2
+# MULTI-NEXT: Metadata: 0x3
# MULTI-NEXT: - Name: '.llvm_bb_addr_map (1)'
# MULTI-NEXT: Type: SHT_LLVM_BB_ADDR_MAP
# MULTI-NEXT: Entries:
# MULTI-NEXT: - Version: 0
-# MULTI-NEXT: Address: 0x20
-# MULTI-NEXT: BBEntries: []
+# MULTI-NEXT: BBRanges:
+# MULTI-NEXT: - BaseAddress: 0x20
+# MULTI-NEXT: BBEntries: []
--- !ELF
FileHeader:
@@ -141,16 +147,18 @@ Sections:
## they are zero.
- Version: 0
Feature: 0x0
- Address: 0x0
- BBEntries:
- - AddressOffset: 0x1
- Size: 0x2
- Metadata: 0x3
+ BBRanges:
+ - BaseAddress: 0x0
+ BBEntries:
+ - AddressOffset: 0x1
+ Size: 0x2
+ Metadata: 0x3
- Name: '.llvm_bb_addr_map (1)'
Type: SHT_LLVM_BB_ADDR_MAP
Entries:
- Version: 0
- Address: 0x20
+ BBRanges:
+ - BaseAddress: 0x20
## Check that obj2yaml uses the "Content" tag to describe an .llvm_bb_addr_map section
## when it can't extract the entries, for example, when the section is truncated, or
@@ -160,7 +168,10 @@ Sections:
# RUN: obj2yaml %t4 | FileCheck %s --check-prefixes=TRUNCATED,INVALID
# RUN: yaml2obj --docnum=1 -DNUMBLOCKS=2 %s -o %t5
-# RUN: obj2yaml %t5 | FileCheck %s --check-prefixes=BADNUMBLOCKS,INVALID
+# RUN: obj2yaml %t5 | FileCheck %s --check-prefixes=BADNUM,INVALID
+
+# RUN: yaml2obj --docnum=1 -DNUMBBRANGES=2 %s -o %t6
+# RUN: obj2yaml %t6 | FileCheck %s --check-prefixes=BADNUM,INVALID
# INVALID: --- !ELF
# INVALID-NEXT: FileHeader:
@@ -170,86 +181,5 @@ Sections:
# INVALID-NEXT: Sections:
# INVALID-NEXT: - Name: .llvm_bb_addr_map
# INVALID-NEXT: Type: SHT_LLVM_BB_ADDR_MAP
-# BADNUMBLOCKS-NEXT: Content: {{([[:xdigit:]]+)}}{{$}}
-# TRUNCATED-NEXT: Content: {{([[:xdigit:]]{16})}}{{$}}
-
-## Check obj2yaml for SHT_LLVM_BB_ADDR_MAP_V0.
-# RUN: yaml2obj --docnum=4 %s -o %t6
-# RUN: obj2yaml %t6 | FileCheck %s --check-prefix=V0
-
-# V0: --- !ELF
-# V0-NEXT: FileHeader:
-# V0-NEXT: Class: ELFCLASS64
-# V0-NEXT: Data: ELFDATA2LSB
-# V0-NEXT: Type: ET_EXEC
-# V0-NEXT: Sections:
-# V0-NEXT: - Name: .llvm_bb_addr_map
-# V0-NEXT: Type: SHT_LLVM_BB_ADDR_MAP_V0
-# V0-NEXT: Entries:
-# V0-NEXT: - Version: 0
-# V0-NEXT: Address: 0x1111
-# V0-NEXT: BBEntries:
-# V0-NEXT: - ID: 0
-# V0-NEXT: AddressOffset: 0x1
-# V0-NEXT: Size: 0x2
-# V0-NEXT: Metadata: 0x3
-
---- !ELF
-FileHeader:
- Class: ELFCLASS64
- Data: ELFDATA2LSB
- Type: ET_EXEC
-Sections:
- - Name: .llvm_bb_addr_map
- Type: SHT_LLVM_BB_ADDR_MAP_V0
- Entries:
- - Version: 0
- Address: 0x1111
- BBEntries:
- - AddressOffset: 0x1
- Size: 0x2
- Metadata: 0x3
-
-## Check obj2yaml for version 1.
-# RUN: yaml2obj --docnum=5 %s -o %t7
-# RUN: obj2yaml %t7 | FileCheck %s --check-prefix=V1
-
-# V1: --- !ELF
-# V1-NEXT: FileHeader:
-# V1-NEXT: Class: ELFCLASS64
-# V1-NEXT: Data: ELFDATA2LSB
-# V1-NEXT: Type: ET_EXEC
-# V1-NEXT: Sections:
-# V1-NEXT: - Name: .llvm_bb_addr_map
-# V1-NEXT: Type: SHT_LLVM_BB_ADDR_MAP
-# V1-NEXT: Entries:
-# V1-NEXT: - Version: 1
-# V1-NEXT: Address: 0x1111
-# V1-NEXT: BBEntries:
-# V1-NEXT: - ID: 0
-# V1-NEXT: AddressOffset: 0x1
-# V1-NEXT: Size: 0x2
-# V1-NEXT: Metadata: 0x3
-# V1-NEXT: - ID: 1
-# V1-NEXT: AddressOffset: 0x4
-# V1-NEXT: Size: 0x5
-# V1-NEXT: Metadata: 0x6
-
---- !ELF
-FileHeader:
- Class: ELFCLASS64
- Data: ELFDATA2LSB
- Type: ET_EXEC
-Sections:
- - Name: .llvm_bb_addr_map
- Type: SHT_LLVM_BB_ADDR_MAP
- Entries:
- - Version: 1
- Address: 0x1111
- BBEntries:
- - AddressOffset: 0x1
- Size: 0x2
- Metadata: 0x3
- - AddressOffset: 0x4
- Size: 0x5
- Metadata: 0x6
+# BADNUM-NEXT: Content: {{([[:xdigit:]]+)}}{{$}}
+# TRUNCATED-NEXT: Content: '{{([[:xdigit:]]{16})}}'{{$}}
diff --git a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
index 2086dc53208b0..74357a4cfa605 100644
--- a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
+++ b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
@@ -86,12 +86,13 @@ Sections:
Type: SHT_LLVM_BB_ADDR_MAP
Entries:
- Version: 2
- Address: 0x0000000000000020
- BBEntries:
- - ID: 11
- AddressOffset: 0x00000001
- Size: 0x00000002
- Metadata: 0x00000003
+ BBRanges:
+ - BaseAddress: 0x0000000000000020
+ BBEntries:
+ - ID: 11
+ AddressOffset: 0x00000001
+ Size: 0x00000002
+ Metadata: 0x00000003
## 5) When specifying the description with Entries, the 'Address' field will be
## zero when omitted.
@@ -99,11 +100,12 @@ Sections:
Type: SHT_LLVM_BB_ADDR_MAP
Entries:
- Version: 2
- BBEntries:
- - ID: 12
- AddressOffset: 0x00000001
- Size: 0x00000002
- Metadata: 0x00000003
+ BBRanges:
+ - BBEntries:
+ - ID: 12
+ AddressOffset: 0x00000001
+ Size: 0x00000002
+ Metadata: 0x00000003
## 6) We can override the NumBlocks field with a value different from the
## actual number of BB Entries.
@@ -111,13 +113,14 @@ Sections:
Type: SHT_LLVM_BB_ADDR_MAP
Entries:
- Version: 2
- Address: 0x0000000000000020
- NumBlocks: 2
- BBEntries:
- - ID: 13
- AddressOffset: 0x00000001
- Size: 0x00000002
- Metadata: 0x00000003
+ BBRanges:
+ - BaseAddress: 0x0000000000000020
+ NumBlocks: 2
+ BBEntries:
+ - ID: 13
+ AddressOffset: 0x00000001
+ Size: 0x00000002
+ Metadata: 0x00000003
## Check we can't use Entries at the same time as either Content or Size.
# RUN: not yaml2obj --docnum=2 -DCONTENT="00" %s 2>&1 | FileCheck %s --check-prefix=INVALID
diff --git a/llvm/tools/llvm-objdump/llvm-objdump.cpp b/llvm/tools/llvm-objdump/llvm-objdump.cpp
index 7f57713e6f946..d0b095c02f507 100644
--- a/llvm/tools/llvm-objdump/llvm-objdump.cpp
+++ b/llvm/tools/llvm-objdump/llvm-objdump.cpp
@@ -172,6 +172,106 @@ class OtoolOptTable : public CommonOptTable {
"Mach-O object file displaying tool") {}
};
+struct BBAddrMapLabel {
+ std::string BlockLabel;
+ std::string PGOAnalysis;
+};
+
+
+// This class represents the BBAddrMap and PGOMap associated with a single
+// function.
+class BBAddrMapFunctionEntry {
+ public:
+ BBAddrMapFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap) : AddrMap(std::move(AddrMap)), PGOMap(std::move(PGOMap)) {}
+
+ const BBAddrMap &getAddrMap() const {
+ return AddrMap;
+ }
+
+ // Returns the PGO string associated with the entry of index `PGOBBEntryIndex`
+ // in `PGOMap`.
+ std::string constructPGOLabelString(size_t PGOBBEntryIndex) const {
+ if (!PGOMap.FeatEnable.hasPGOAnalysis())
+ return "";
+ std::string PGOString;
+ raw_string_ostream PGOSS(PGOString);
+
+ PGOSS << " (";
+ if (PGOMap.FeatEnable.FuncEntryCount && PGOBBEntryIndex == 0) {
+ PGOSS << "Entry count: " << Twine(PGOMap.FuncEntryCount);
+ if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
+ PGOSS << ", ";
+ }
+ }
+
+ if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
+
+ assert(PGOBBEntryIndex < PGOMap.BBEntries.size() &&
+ "Expected PGOAnalysisMap and BBAddrMap to have the same entires");
+ const PGOAnalysisMap::PGOBBEntry &PGOBBEntry =
+ PGOMap.BBEntries[PGOBBEntryIndex];
+
+ if (PGOMap.FeatEnable.BBFreq) {
+ PGOSS << "Frequency: " << Twine(PGOBBEntry.BlockFreq.getFrequency());
+ if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
+ PGOSS << ", ";
+ }
+ }
+ if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
+ PGOSS << "Successors: ";
+ interleaveComma(
+ PGOBBEntry.Successors, PGOSS,
+ [&PGOSS](const PGOAnalysisMap::PGOBBEntry::SuccessorEntry &SE) {
+ PGOSS << "BB" << SE.ID << ":";
+ PGOSS.write_hex(SE.Prob.getNumerator());
+ });
+ }
+ }
+ PGOSS << ")";
+
+ return PGOString;
+}
+private:
+ const BBAddrMap AddrMap;
+ const PGOAnalysisMap PGOMap;
+
+ };
+
+// This class represents the BBAddrMap and PGOMap of potentially multiple
+// functions in a section.
+class BBAddrMapInfo {
+ public:
+ void clear() {
+ FunctionAddrToMap.clear();
+ RangeBaseAddrToFunctionAddr.clear();
+ }
+ bool empty() const {
+ return FunctionAddrToMap.empty();
+ }
+
+ void AddFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap) {
+ uint64_t FunctionAddr = AddrMap.getFunctionAddress();
+ for (size_t I=1; I < AddrMap.BBRanges.size(); ++I)
+ RangeBaseAddrToFunctionAddr.emplace(AddrMap.BBRanges[I].BaseAddress, FunctionAddr);
+ [[maybe_unused]] auto R = FunctionAddrToMap.try_emplace(FunctionAddr, std::move(AddrMap), std::move(PGOMap));
+ assert(R.second && "duplicate function address");
+ }
+
+ const BBAddrMapFunctionEntry *getEntryForAddress(uint64_t BaseAddress) const {
+ uint64_t FunctionAddr = BaseAddress;
+ auto S = RangeBaseAddrToFunctionAddr.find(BaseAddress);
+ if (S != RangeBaseAddrToFunctionAddr.end())
+ FunctionAddr = S->second;
+ auto R = FunctionAddrToMap.find(FunctionAddr);
+ if (R == FunctionAddrToMap.end()) return nullptr;
+ return &R->second;
+ }
+
+ private:
+ std::unordered_map<uint64_t, BBAddrMapFunctionEntry> FunctionAddrToMap;
+ std::unordered_map<uint64_t, uint64_t> RangeBaseAddrToFunctionAddr;
+};
+
} // namespace
#define DEBUG_TYPE "objdump"
@@ -1266,80 +1366,33 @@ static SymbolInfoTy createDummySymbolInfo(const ObjectFile &Obj,
return SymbolInfoTy(Addr, Name, Type);
}
-struct BBAddrMapLabel {
- std::string BlockLabel;
- std::string PGOAnalysis;
-};
-
-static std::string constructPGOLabelString(const PGOAnalysisMap &PGOMap,
- size_t BBEntryIndex) {
- std::string PGOString;
- raw_string_ostream PGOSS(PGOString);
-
- PGOSS << " (";
- if (PGOMap.FeatEnable.FuncEntryCount && BBEntryIndex == 0) {
- PGOSS << "Entry count: " << Twine(PGOMap.FuncEntryCount);
- if (PGOMap.FeatEnable.BBFreq || PGOMap.FeatEnable.BrProb) {
- PGOSS << ", ";
- }
- }
-
- if (PGOMap.FeatEnable.BBFreq || PGOMap.FeatEnable.BrProb) {
- assert(BBEntryIndex < PGOMap.BBEntries.size() &&
- "Expected PGOAnalysisMap and BBAddrMap to have the same entires");
- const PGOAnalysisMap::PGOBBEntry &PGOBBEntry =
- PGOMap.BBEntries[BBEntryIndex];
-
- if (PGOMap.FeatEnable.BBFreq) {
- PGOSS << "Frequency: " << Twine(PGOBBEntry.BlockFreq.getFrequency());
- if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
- PGOSS << ", ";
- }
- }
- if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
- PGOSS << "Successors: ";
- interleaveComma(
- PGOBBEntry.Successors, PGOSS,
- [&PGOSS](const PGOAnalysisMap::PGOBBEntry::SuccessorEntry &SE) {
- PGOSS << "BB" << SE.ID << ":";
- PGOSS.write_hex(SE.Prob.getNumerator());
- });
- }
- }
- PGOSS << ")";
-
- return PGOString;
-}
-
static void collectBBAddrMapLabels(
- const std::unordered_map<uint64_t, BBAddrMap> &AddrToBBAddrMap,
- const std::unordered_map<uint64_t, PGOAnalysisMap> &AddrToPGOAnalysisMap,
+ const BBAddrMapInfo &FullAddrMap,
uint64_t SectionAddr, uint64_t Start, uint64_t End,
- std::unordered_map<uint64_t, std::vector<BBAddrMapLabel>> &Labels,
- const StringRef FileName) {
- if (AddrToBBAddrMap.empty())
+ std::unordered_map<uint64_t, std::vector<BBAddrMapLabel>> &Labels) {
+ if (FullAddrMap.empty())
return;
Labels.clear();
uint64_t StartAddress = SectionAddr + Start;
uint64_t EndAddress = SectionAddr + End;
- auto Iter = AddrToBBAddrMap.find(StartAddress);
- if (Iter == AddrToBBAddrMap.end())
+ const BBAddrMapFunctionEntry *FunctionMap = FullAddrMap.getEntryForAddress(StartAddress);
+ if (!FunctionMap)
return;
- auto PGOIter = AddrToPGOAnalysisMap.find(StartAddress);
-
- for (size_t I = 0; I < Iter->second.getBBEntries().size(); ++I) {
- const BBAddrMap::BBEntry &BBEntry = Iter->second.getBBEntries()[I];
- uint64_t BBAddress = BBEntry.Offset + Iter->second.getFunctionAddress();
+ std::optional<size_t> BBRangeIndex = FunctionMap->getAddrMap().getBBRangeIndexForBaseAddress(StartAddress);
+ if (!BBRangeIndex)
+ return;
+ size_t NumBBEntriesBeforeRange = 0;
+ for (size_t I=0; I < *BBRangeIndex; ++I)
+ NumBBEntriesBeforeRange += FunctionMap->getAddrMap().BBRanges[I].BBEntries.size();
+ const auto &BBRange = FunctionMap->getAddrMap().BBRanges[*BBRangeIndex];
+ for (size_t I = 0; I < BBRange.BBEntries.size(); ++I) {
+ const BBAddrMap::BBEntry &BBEntry = BBRange.BBEntries[I];
+ uint64_t BBAddress = BBEntry.Offset + BBRange.BaseAddress;
if (BBAddress >= EndAddress)
continue;
std::string LabelString = ("BB" + Twine(BBEntry.ID)).str();
- std::string PGOString;
-
- if (PGOIter != AddrToPGOAnalysisMap.end())
- PGOString = constructPGOLabelString(PGOIter->second, I);
-
- Labels[BBAddress].push_back({LabelString, PGOString});
+ Labels[BBAddress].push_back({LabelString, FunctionMap->constructPGOLabelString(NumBBEntriesBeforeRange + I)});
}
}
@@ -1696,11 +1749,10 @@ disassembleObject(ObjectFile &Obj, const ObjectFile &DbgObj,
LLVM_DEBUG(LVP.dump());
- std::unordered_map<uint64_t, BBAddrMap> AddrToBBAddrMap;
- std::unordered_map<uint64_t, PGOAnalysisMap> AddrToPGOAnalysisMap;
+ BBAddrMapInfo FullAddrMap;
auto ReadBBAddrMap = [&](std::optional<unsigned> SectionIndex =
std::nullopt) {
- AddrToBBAddrMap.clear();
+ FullAddrMap.clear();
if (const auto *Elf = dyn_cast<ELFObjectFileBase>(&Obj)) {
std::vector<PGOAnalysisMap> PGOAnalyses;
auto BBAddrMapsOrErr = Elf->readBBAddrMap(SectionIndex, &PGOAnalyses);
@@ -1710,10 +1762,7 @@ disassembleObject(ObjectFile &Obj, const ObjectFile &DbgObj,
}
for (auto &&[FunctionBBAddrMap, FunctionPGOAnalysis] :
zip_equal(*std::move(BBAddrMapsOrErr), std::move(PGOAnalyses))) {
- uint64_t Addr = FunctionBBAddrMap.Addr;
- AddrToBBAddrMap.emplace(Addr, std::move(FunctionBBAddrMap));
- if (FunctionPGOAnalysis.FeatEnable.anyEnabled())
- AddrToPGOAnalysisMap.emplace(Addr, std::move(FunctionPGOAnalysis));
+ FullAddrMap.AddFunctionEntry(std::move(FunctionBBAddrMap), std::move(FunctionPGOAnalysis));
}
}
};
@@ -2049,9 +2098,8 @@ disassembleObject(ObjectFile &Obj, const ObjectFile &DbgObj,
DT->DisAsm.get(), DT->InstPrinter.get(),
PrimaryTarget.SubtargetInfo.get(),
SectionAddr, Index, End, AllLabels);
- collectBBAddrMapLabels(AddrToBBAddrMap, AddrToPGOAnalysisMap,
- SectionAddr, Index, End, BBAddrMapLabels,
- FileName);
+ collectBBAddrMapLabels(FullAddrMap,
+ SectionAddr, Index, End, BBAddrMapLabels);
}
if (DT->InstrAnalysis)
diff --git a/llvm/tools/llvm-readobj/ELFDumper.cpp b/llvm/tools/llvm-readobj/ELFDumper.cpp
index 4bf4640f1a4ca..1005c81ff8fdf 100644
--- a/llvm/tools/llvm-readobj/ELFDumper.cpp
+++ b/llvm/tools/llvm-readobj/ELFDumper.cpp
@@ -7545,8 +7545,7 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
bool IsRelocatable = this->Obj.getHeader().e_type == ELF::ET_REL;
using Elf_Shdr = typename ELFT::Shdr;
auto IsMatch = [](const Elf_Shdr &Sec) -> bool {
- return Sec.sh_type == ELF::SHT_LLVM_BB_ADDR_MAP ||
- Sec.sh_type == ELF::SHT_LLVM_BB_ADDR_MAP_V0;
+ return Sec.sh_type == ELF::SHT_LLVM_BB_ADDR_MAP;
};
Expected<MapVector<const Elf_Shdr *, const Elf_Shdr *>> SecRelocMapOrErr =
this->Obj.getSectionAndRelocations(IsMatch);
@@ -7575,7 +7574,7 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
": " + toString(BBAddrMapOrErr.takeError()));
continue;
}
- for (const auto &[AM, PAM] : zip_equal(*BBAddrMapOrErr, PGOAnalyses)) {
+ for (const BBAddrMap &AM : *BBAddrMapOrErr) {
DictScope D(W, "Function");
W.printHex("At", AM.Addr);
SmallVector<uint32_t> FuncSymIndex =
@@ -7588,11 +7587,13 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
else
FuncName = this->getStaticSymbolName(FuncSymIndex.front());
W.printString("Name", FuncName);
-
- {
- ListScope L(W, "BB entries");
- for (const BBAddrMap::BBEntry &BBE : AM.BBEntries) {
- DictScope L(W);
+ ListScope BBRL(W, "BB Ranges");
+ for (const BBAddrMap::BBRangeEntry &BBR : AM.BBRanges) {
+ DictScope BBRD(W);
+ W.printHex("Base Address", BBR.BaseAddress);
+ ListScope BBEL(W, "BB Entries");
+ for (const BBAddrMap::BBEntry &BBE : BBR.BBEntries) {
+ DictScope BBED(W);
W.printNumber("ID", BBE.ID);
W.printHex("Offset", BBE.Offset);
W.printHex("Size", BBE.Size);
diff --git a/llvm/tools/obj2yaml/elf2yaml.cpp b/llvm/tools/obj2yaml/elf2yaml.cpp
index 86bd138c22064..38a69f77ff836 100644
--- a/llvm/tools/obj2yaml/elf2yaml.cpp
+++ b/llvm/tools/obj2yaml/elf2yaml.cpp
@@ -626,7 +626,6 @@ ELFDumper<ELFT>::dumpSections() {
case ELF::SHT_LLVM_CALL_GRAPH_PROFILE:
return
[this](const Elf_Shdr *S) { return dumpCallGraphProfileSection(S); };
- case ELF::SHT_LLVM_BB_ADDR_MAP_V0:
case ELF::SHT_LLVM_BB_ADDR_MAP:
return [this](const Elf_Shdr *S) { return dumpBBAddrMapSection(S); };
case ELF::SHT_STRTAB:
@@ -893,6 +892,7 @@ ELFDumper<ELFT>::dumpBBAddrMapSection(const Elf_Shdr *Shdr) {
DataExtractor::Cursor Cur(0);
uint8_t Version = 0;
uint8_t Feature = 0;
+ uint64_t Address = 0;
while (Cur && Cur.tell() < Content.size()) {
if (Shdr->sh_type == ELF::SHT_LLVM_BB_ADDR_MAP) {
Version = Data.getU8(Cur);
@@ -903,19 +903,41 @@ ELFDumper<ELFT>::dumpBBAddrMapSection(const Elf_Shdr *Shdr) {
Twine(static_cast<int>(Version)));
Feature = Data.getU8(Cur);
}
- uint64_t Address = Data.getAddress(Cur);
- uint64_t NumBlocks = Data.getULEB128(Cur);
- std::vector<ELFYAML::BBAddrMapEntry::BBEntry> BBEntries;
- // Read the specified number of BB entries, or until decoding fails.
- for (uint64_t BlockIndex = 0; Cur && BlockIndex < NumBlocks; ++BlockIndex) {
- uint32_t ID = Version >= 2 ? Data.getULEB128(Cur) : BlockIndex;
- uint64_t Offset = Data.getULEB128(Cur);
- uint64_t Size = Data.getULEB128(Cur);
- uint64_t Metadata = Data.getULEB128(Cur);
- BBEntries.push_back({ID, Offset, Size, Metadata});
+ uint64_t NumBBRanges = 1;
+ uint64_t NumBlocks = 0;
+ auto FeatureOrErr = llvm::object::BBAddrMap::Features::decode(Feature);
+ if (!FeatureOrErr)
+ return FeatureOrErr.takeError();
+ if (FeatureOrErr->MultiBBRange) {
+ NumBBRanges = Data.getULEB128(Cur);
+ } else {
+ Address = Data.getAddress(Cur);
+ NumBlocks = Data.getULEB128(Cur);
+ }
+ std::vector<ELFYAML::BBAddrMapEntry::BBRangeEntry> BBRanges;
+ uint64_t BaseAddress = 0;
+ for (uint64_t BBRangeN = 0; Cur && BBRangeN != NumBBRanges; ++BBRangeN) {
+ if (FeatureOrErr->MultiBBRange) {
+ BaseAddress = Data.getAddress(Cur);
+ NumBlocks = Data.getULEB128(Cur);
+ } else {
+ BaseAddress = Address;
+ }
+
+ std::vector<ELFYAML::BBAddrMapEntry::BBEntry> BBEntries;
+ // Read the specified number of BB entries, or until decoding fails.
+ for (uint64_t BlockIndex = 0; Cur && BlockIndex < NumBlocks;
+ ++BlockIndex) {
+ uint32_t ID = Version >= 2 ? Data.getULEB128(Cur) : BlockIndex;
+ uint64_t Offset = Data.getULEB128(Cur);
+ uint64_t Size = Data.getULEB128(Cur);
+ uint64_t Metadata = Data.getULEB128(Cur);
+ BBEntries.push_back({ID, Offset, Size, Metadata});
+ }
+ BBRanges.push_back({BaseAddress, /*NumBlocks=*/{}, BBEntries});
}
Entries.push_back(
- {Version, Feature, Address, /*NumBlocks=*/{}, std::move(BBEntries)});
+ {Version, Feature, /*NumBBRanges=*/{}, std::move(BBRanges)});
}
if (!Cur) {
diff --git a/llvm/unittests/Object/ELFObjectFileTest.cpp b/llvm/unittests/Object/ELFObjectFileTest.cpp
index c539d79ac14e8..99817913ca1be 100644
--- a/llvm/unittests/Object/ELFObjectFileTest.cpp
+++ b/llvm/unittests/Object/ELFObjectFileTest.cpp
@@ -511,7 +511,6 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
- Type: SHT_LLVM_BB_ADDR_MAP
Name: .llvm_bb_addr_map
Entries:
- - Address: 0x11111
)");
auto DoCheck = [&](StringRef YamlString, const char *ErrMsg) {
@@ -531,11 +530,13 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
// Check that we can detect unsupported versions.
SmallString<128> UnsupportedVersionYamlString(CommonYamlString);
UnsupportedVersionYamlString += R"(
- Version: 3
- BBEntries:
- - AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ - Version: 3
+ BBRanges:
+ - BaseAddress: 0x11111
+ BBEntries:
+ - AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
)";
{
@@ -546,12 +547,13 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
SmallString<128> CommonVersionedYamlString(CommonYamlString);
CommonVersionedYamlString += R"(
- Version: 2
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ - Version: 2
+ BBRanges:
+ - BaseAddress: 0x11111
+ BBEntries:
+ - AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
)";
// Check that we can detect the malformed encoding when the section is
@@ -572,24 +574,24 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
SmallVector<SmallString<128>, 3> OverInt32LimitYamlStrings(
3, CommonVersionedYamlString);
OverInt32LimitYamlStrings[0] += R"(
- - ID: 1
- AddressOffset: 0x100000000
- Size: 0xFFFFFFFF
- Metadata: 0xFFFFFFFF
+ - ID: 1
+ AddressOffset: 0x100000000
+ Size: 0xFFFFFFFF
+ Metadata: 0xFFFFFFFF
)";
OverInt32LimitYamlStrings[1] += R"(
- - ID: 2
- AddressOffset: 0xFFFFFFFF
- Size: 0x100000000
- Metadata: 0xFFFFFFFF
+ - ID: 2
+ AddressOffset: 0xFFFFFFFF
+ Size: 0x100000000
+ Metadata: 0xFFFFFFFF
)";
OverInt32LimitYamlStrings[2] += R"(
- - ID: 3
- AddressOffset: 0xFFFFFFFF
- Size: 0xFFFFFFFF
- Metadata: 0x100000000
+ - ID: 3
+ AddressOffset: 0xFFFFFFFF
+ Size: 0xFFFFFFFF
+ Metadata: 0x100000000
)";
{
@@ -635,7 +637,7 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
// with an out-of-range value.
SmallString<128> OverLimitNumBlocks(CommonVersionedYamlString);
OverLimitNumBlocks += R"(
- NumBlocks: 0x100000000
+ NumBlocks: 0x100000000
)";
{
@@ -643,6 +645,16 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
DoCheck(OverLimitNumBlocks,
"ULEB128 value at offset 0xa exceeds UINT32_MAX (0x100000000)");
}
+
+ // Check for proper error handling when the 'NumBBRanges' field is overridden
+ // with an out-of-range value.
+ SmallString<128> OverLimitNumBBRanges(CommonVersionedYamlString);
+ OverLimitNumBBRanges += R"(
+ NumBBRanges: 0x100000000
+ Feature: 0x8
+)";
+ DoCheck(OverLimitNumBBRanges,
+ "ULEB128 value at offset 0x2 exceeds UINT32_MAX (0x100000000)");
}
// Test for the ELFObjectFile::readBBAddrMap API.
@@ -659,51 +671,67 @@ TEST(ELFObjectFileTest, ReadBBAddrMap) {
Link: 1
Entries:
- Version: 2
- Address: 0x11111
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: 0x11111
+ BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
- Name: .llvm_bb_addr_map_2
Type: SHT_LLVM_BB_ADDR_MAP
Link: 1
Entries:
- Version: 2
- Address: 0x22222
- BBEntries:
- - ID: 2
- AddressOffset: 0x0
- Size: 0x2
- Metadata: 0x4
+ Feature: 0x8
+ BBRanges:
+ - BaseAddress: 0x22222
+ BBEntries:
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x2
+ Metadata: 0x4
+ - BaseAddress: 0xFFFFF
+ BBEntries:
+ - ID: 15
+ AddressOffset: 0xF0
+ Size: 0xF1
+ Metadata: 0x1F
- Name: .llvm_bb_addr_map_3
Type: SHT_LLVM_BB_ADDR_MAP
Link: 2
Entries:
- Version: 1
- Address: 0x33333
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x3
- Metadata: 0x6
+ BBRanges:
+ - BaseAddress: 0x33333
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x3
+ Metadata: 0x6
- Name: .llvm_bb_addr_map_4
- Type: SHT_LLVM_BB_ADDR_MAP_V0
+ Type: SHT_LLVM_BB_ADDR_MAP
# Link: 0 (by default, can be overriden)
Entries:
- - Version: 0
- Address: 0x44444
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x4
- Metadata: 0x18
+ - Version: 2
+ BBRanges:
+ - BaseAddress: 0x44444
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x4
+ Metadata: 0x18
)");
- BBAddrMap E1(0x11111, {{1, 0x0, 0x1, {false, true, false, false, false}}});
- BBAddrMap E2(0x22222, {{2, 0x0, 0x2, {false, false, true, false, false}}});
- BBAddrMap E3(0x33333, {{0, 0x0, 0x3, {false, true, true, false, false}}});
- BBAddrMap E4(0x44444, {{0, 0x0, 0x4, {false, false, false, true, true}}});
+ BBAddrMap E1 = {
+ {{0x11111, {{1, 0x0, 0x1, {false, true, false, false, false}}}}}};
+ BBAddrMap E2 = {
+ {{0x22222, {{2, 0x0, 0x2, {false, false, true, false, false}}}},
+ {0xFFFFF, {{15, 0xF0, 0xF1, {true, true, true, true, true}}}}}};
+ BBAddrMap E3 = {
+ {{0x33333, {{0, 0x0, 0x3, {false, true, true, false, false}}}}}};
+ BBAddrMap E4 = {
+ {{0x44444, {{0, 0x0, 0x4, {false, false, false, true, true}}}}}};
std::vector<BBAddrMap> Section0BBAddrMaps = {E4};
std::vector<BBAddrMap> Section1BBAddrMaps = {E3};
@@ -767,13 +795,13 @@ TEST(ELFObjectFileTest, ReadBBAddrMap) {
// (not present) section.
SmallString<128> InvalidLinkedYamlString(CommonYamlString);
InvalidLinkedYamlString += R"(
- Link: 10
+ Link: 121
)";
DoCheckFails(InvalidLinkedYamlString, /*TextSectionIndex=*/4,
"unable to get the linked-to section for "
- "SHT_LLVM_BB_ADDR_MAP_V0 section with index 4: invalid section "
- "index: 10");
+ "SHT_LLVM_BB_ADDR_MAP section with index 4: invalid section "
+ "index: 121");
{
SCOPED_TRACE("invalid linked section");
// Linked sections are not checked when we don't target a specific text
@@ -785,14 +813,14 @@ TEST(ELFObjectFileTest, ReadBBAddrMap) {
// Check that we can detect when bb-address-map decoding fails.
SmallString<128> TruncatedYamlString(CommonYamlString);
TruncatedYamlString += R"(
- ShSize: 0x8
+ ShSize: 0xa
)";
{
SCOPED_TRACE("truncated section");
DoCheckFails(TruncatedYamlString, /*TextSectionIndex=*/std::nullopt,
- "unable to read SHT_LLVM_BB_ADDR_MAP_V0 section with index 4: "
- "unable to decode LEB128 at offset 0x00000008: malformed "
+ "unable to read SHT_LLVM_BB_ADDR_MAP section with index 4: "
+ "unable to decode LEB128 at offset 0x0000000a: malformed "
"uleb128, extends past end");
// Check that we can read the other section's bb-address-maps which are
@@ -815,7 +843,6 @@ TEST(ELFObjectFileTest, InvalidDecodePGOAnalysisMap) {
- Type: SHT_LLVM_BB_ADDR_MAP
Name: .llvm_bb_addr_map
Entries:
- - Address: 0x11111
)");
auto DoCheck = [&](StringRef YamlString, const char *ErrMsg) {
@@ -839,12 +866,13 @@ TEST(ELFObjectFileTest, InvalidDecodePGOAnalysisMap) {
// Check that we can detect unsupported versions that are too old.
SmallString<128> UnsupportedLowVersionYamlString(CommonYamlString);
UnsupportedLowVersionYamlString += R"(
- Version: 1
+ - Version: 1
Feature: 0x4
- BBEntries:
- - AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BBEntries:
+ - AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
)";
{
@@ -856,38 +884,39 @@ TEST(ELFObjectFileTest, InvalidDecodePGOAnalysisMap) {
SmallString<128> CommonVersionedYamlString(CommonYamlString);
CommonVersionedYamlString += R"(
- Version: 2
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ - Version: 2
+ BBRanges:
+ - BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
)";
// Check that we fail when function entry count is enabled but not provided.
SmallString<128> MissingFuncEntryCount(CommonYamlString);
MissingFuncEntryCount += R"(
- Version: 2
+ - Version: 2
Feature: 0x01
)";
{
SCOPED_TRACE("missing function entry count");
DoCheck(MissingFuncEntryCount,
- "unable to decode LEB128 at offset 0x0000000b: malformed uleb128, "
- "extends past end");
+ "unexpected end of data at offset 0x2 while reading [0x2, 0xa)");
}
// Check that we fail when basic block frequency is enabled but not provided.
SmallString<128> MissingBBFreq(CommonYamlString);
MissingBBFreq += R"(
- Version: 2
+ - Version: 2
Feature: 0x02
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
)";
{
@@ -899,21 +928,22 @@ TEST(ELFObjectFileTest, InvalidDecodePGOAnalysisMap) {
// Check that we fail when branch probability is enabled but not provided.
SmallString<128> MissingBrProb(CommonYamlString);
MissingBrProb += R"(
- Version: 2
+ - Version: 2
Feature: 0x04
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x6
- - ID: 2
- AddressOffset: 0x1
- Size: 0x1
- Metadata: 0x2
- - ID: 3
- AddressOffset: 0x2
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x6
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x1
+ Metadata: 0x2
+ - ID: 3
+ AddressOffset: 0x2
+ Size: 0x1
+ Metadata: 0x2
PGOAnalyses:
- PGOBBEntries:
- Successors:
@@ -947,13 +977,14 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
Link: 1
Entries:
- Version: 2
- Address: 0x11111
Feature: 0x1
- BBEntries:
- - ID: 1
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ BBRanges:
+ - BaseAddress: 0x11111
+ BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
PGOAnalyses:
- FuncEntryCount: 892
- Name: .llvm_bb_addr_map_2
@@ -961,13 +992,14 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
Link: 1
Entries:
- Version: 2
- Address: 0x22222
Feature: 0x2
- BBEntries:
- - ID: 2
- AddressOffset: 0x0
- Size: 0x2
- Metadata: 0x4
+ BBRanges:
+ - BaseAddress: 0x22222
+ BBEntries:
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x2
+ Metadata: 0x4
PGOAnalyses:
- PGOBBEntries:
- BBFreq: 343
@@ -976,21 +1008,22 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
Link: 2
Entries:
- Version: 2
- Address: 0x33333
Feature: 0x4
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x3
- Metadata: 0x6
- - ID: 1
- AddressOffset: 0x0
- Size: 0x3
- Metadata: 0x4
- - ID: 2
- AddressOffset: 0x0
- Size: 0x3
- Metadata: 0x0
+ BBRanges:
+ - BaseAddress: 0x33333
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x3
+ Metadata: 0x6
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x3
+ Metadata: 0x4
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x3
+ Metadata: 0x0
PGOAnalyses:
- PGOBBEntries:
- Successors:
@@ -1007,25 +1040,26 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
# Link: 0 (by default, can be overriden)
Entries:
- Version: 2
- Address: 0x44444
Feature: 0x7
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x4
- Metadata: 0x18
- - ID: 1
- AddressOffset: 0x0
- Size: 0x4
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x0
- Size: 0x4
- Metadata: 0x0
- - ID: 3
- AddressOffset: 0x0
- Size: 0x4
- Metadata: 0x0
+ BBRanges:
+ - BaseAddress: 0x44444
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x4
+ Metadata: 0x18
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x4
+ Metadata: 0x0
PGOAnalyses:
- FuncEntryCount: 1000
PGOBBEntries:
@@ -1054,34 +1088,74 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
# Link: 0 (by default, can be overriden)
Entries:
- Version: 2
- Address: 0x55555
Feature: 0x0
- BBEntries:
- - ID: 2
- AddressOffset: 0x0
- Size: 0x2
- Metadata: 0x4
+ BBRanges:
+ - BaseAddress: 0x55555
+ BBEntries:
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x2
+ Metadata: 0x4
PGOAnalyses: [{}]
- )");
-
- BBAddrMap E1(0x11111, {{1, 0x0, 0x1, {false, true, false, false, false}}});
- PGOAnalysisMap P1 = {892, {}, {true, false, false}};
- BBAddrMap E2(0x22222, {{2, 0x0, 0x2, {false, false, true, false, false}}});
- PGOAnalysisMap P2 = {{}, {{BlockFrequency(343), {}}}, {false, true, false}};
- BBAddrMap E3(0x33333, {{0, 0x0, 0x3, {false, true, true, false, false}},
- {1, 0x3, 0x3, {false, false, true, false, false}},
- {2, 0x6, 0x3, {false, false, false, false, false}}});
+ - Name: .llvm_bb_addr_map_6
+ Type: SHT_LLVM_BB_ADDR_MAP
+ # Link: 0 (by default, can be overriden)
+ Entries:
+ - Version: 2
+ Feature: 0xc
+ BBRanges:
+ - BaseAddress: 0x66666
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x6
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x4
+ - BaseAddress: 0x666661
+ BBEntries:
+ - ID: 2
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ PGOAnalyses:
+ - PGOBBEntries:
+ - Successors:
+ - ID: 1
+ BrProb: 0x22222222
+ - ID: 2
+ BrProb: 0xcccccccc
+ - Successors:
+ - ID: 2
+ BrProb: 0x88888888
+ - Successors: []
+)");
+
+ BBAddrMap E1 = {
+ {{0x11111, {{1, 0x0, 0x1, {false, true, false, false, false}}}}}};
+ PGOAnalysisMap P1 = {892, {}, {true, false, false, false}};
+ BBAddrMap E2 = {
+ {{0x22222, {{2, 0x0, 0x2, {false, false, true, false, false}}}}}};
+ PGOAnalysisMap P2 = {
+ {}, {{BlockFrequency(343), {}}}, {false, true, false, false}};
+ BBAddrMap E3 = {{{0x33333,
+ {{0, 0x0, 0x3, {false, true, true, false, false}},
+ {1, 0x3, 0x3, {false, false, true, false, false}},
+ {2, 0x6, 0x3, {false, false, false, false, false}}}}}};
PGOAnalysisMap P3 = {{},
{{{},
{{1, BranchProbability::getRaw(0x1111'1111)},
{2, BranchProbability::getRaw(0xeeee'eeee)}}},
{{}, {{2, BranchProbability::getRaw(0xffff'ffff)}}},
{{}, {}}},
- {false, false, true}};
- BBAddrMap E4(0x44444, {{0, 0x0, 0x4, {false, false, false, true, true}},
- {1, 0x4, 0x4, {false, false, false, false, false}},
- {2, 0x8, 0x4, {false, false, false, false, false}},
- {3, 0xc, 0x4, {false, false, false, false, false}}});
+ {false, false, true, false}};
+ BBAddrMap E4 = {{{0x44444,
+ {{0, 0x0, 0x4, {false, false, false, true, true}},
+ {1, 0x4, 0x4, {false, false, false, false, false}},
+ {2, 0x8, 0x4, {false, false, false, false, false}},
+ {3, 0xc, 0x4, {false, false, false, false, false}}}}}};
PGOAnalysisMap P4 = {
1000,
{{BlockFrequency(1000),
@@ -1093,19 +1167,32 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
{3, BranchProbability::getRaw(0xeeee'eeee)}}},
{BlockFrequency(18), {{3, BranchProbability::getRaw(0xffff'ffff)}}},
{BlockFrequency(1000), {}}},
- {true, true, true}};
- BBAddrMap E5(0x55555, {{2, 0x0, 0x2, {false, false, true, false, false}}});
- PGOAnalysisMap P5 = {{}, {}, {false, false, false}};
+ {true, true, true, false}};
+ BBAddrMap E5 = {
+ {{0x55555, {{2, 0x0, 0x2, {false, false, true, false, false}}}}}};
+ PGOAnalysisMap P5 = {{}, {}, {false, false, false, false}};
+ BBAddrMap E6 = {
+ {{0x66666,
+ {{0, 0x0, 0x6, {false, true, true, false, false}},
+ {1, 0x6, 0x6, {false, false, true, false, false}}}},
+ {0x666661, {{2, 0x0, 0x6, {false, false, false, false, false}}}}}};
+ PGOAnalysisMap P6 = {{},
+ {{{},
+ {{1, BranchProbability::getRaw(0x2222'2222)},
+ {2, BranchProbability::getRaw(0xcccc'cccc)}}},
+ {{}, {{2, BranchProbability::getRaw(0x8888'8888)}}},
+ {{}, {}}},
+ {false, false, true, true}};
- std::vector<BBAddrMap> Section0BBAddrMaps = {E4, E5};
+ std::vector<BBAddrMap> Section0BBAddrMaps = {E4, E5, E6};
std::vector<BBAddrMap> Section1BBAddrMaps = {E3};
std::vector<BBAddrMap> Section2BBAddrMaps = {E1, E2};
- std::vector<BBAddrMap> AllBBAddrMaps = {E1, E2, E3, E4, E5};
+ std::vector<BBAddrMap> AllBBAddrMaps = {E1, E2, E3, E4, E5, E6};
- std::vector<PGOAnalysisMap> Section0PGOAnalysisMaps = {P4, P5};
+ std::vector<PGOAnalysisMap> Section0PGOAnalysisMaps = {P4, P5, P6};
std::vector<PGOAnalysisMap> Section1PGOAnalysisMaps = {P3};
std::vector<PGOAnalysisMap> Section2PGOAnalysisMaps = {P1, P2};
- std::vector<PGOAnalysisMap> AllPGOAnalysisMaps = {P1, P2, P3, P4, P5};
+ std::vector<PGOAnalysisMap> AllPGOAnalysisMaps = {P1, P2, P3, P4, P5, P6};
auto DoCheckSucceeds =
[&](StringRef YamlString, std::optional<unsigned> TextSectionIndex,
@@ -1131,10 +1218,21 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
EXPECT_EQ(*BBAddrMaps, ExpectedResult);
if (ExpectedPGO) {
EXPECT_EQ(BBAddrMaps->size(), PGOAnalyses.size());
+ for (const auto &PGO : PGOAnalyses) {
+ errs() << "FuncEntryCount: " << PGO.FuncEntryCount << "\n";
+ for (const auto &PGOBB : PGO.BBEntries)
+ errs() << "\tBB: " << PGOBB.BlockFreq.getFrequency() << "\n";
+ }
+ errs() << " To expected:\n";
+ for (const auto &PGO : *ExpectedPGO) {
+ errs() << "FuncEntryCount: " << PGO.FuncEntryCount << "\n";
+ for (const auto &PGOBB : PGO.BBEntries)
+ errs() << "\tBB: " << PGOBB.BlockFreq.getFrequency() << "\n";
+ }
EXPECT_EQ(PGOAnalyses, *ExpectedPGO);
for (auto &&[BB, PGO] : llvm::zip(*BBAddrMaps, PGOAnalyses)) {
if (PGO.FeatEnable.BBFreq || PGO.FeatEnable.BrProb)
- EXPECT_EQ(BB.getBBEntries().size(), PGO.BBEntries.size());
+ EXPECT_EQ(BB.getNumBBEntries(), PGO.BBEntries.size());
}
}
};
@@ -1190,15 +1288,15 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
// (not present) section.
SmallString<128> InvalidLinkedYamlString(CommonYamlString);
InvalidLinkedYamlString += R"(
- Link: 10
+ Link: 121
)";
{
SCOPED_TRACE("invalid linked section");
DoCheckFails(InvalidLinkedYamlString, /*TextSectionIndex=*/5,
"unable to get the linked-to section for "
- "SHT_LLVM_BB_ADDR_MAP section with index 5: invalid section "
- "index: 10");
+ "SHT_LLVM_BB_ADDR_MAP section with index 6: invalid section "
+ "index: 121");
// Linked sections are not checked when we don't target a specific text
// section.
@@ -1216,10 +1314,10 @@ TEST(ELFObjectFileTest, ReadPGOAnalysisMap) {
{
SCOPED_TRACE("truncated section");
- DoCheckFails(TruncatedYamlString, /*TextSectionIndex=*/std::nullopt,
- "unable to read SHT_LLVM_BB_ADDR_MAP section with index 5: "
- "unable to decode LEB128 at offset 0x0000000a: malformed "
- "uleb128, extends past end");
+ DoCheckFails(
+ TruncatedYamlString, /*TextSectionIndex=*/std::nullopt,
+ "unable to read SHT_LLVM_BB_ADDR_MAP section with index 6: "
+ "unexpected end of data at offset 0xa while reading [0x3, 0xb)");
// Check that we can read the other section's bb-address-maps which are
// valid.
DoCheckSucceeds(TruncatedYamlString, /*TextSectionIndex=*/2,
diff --git a/llvm/unittests/Object/ELFTypesTest.cpp b/llvm/unittests/Object/ELFTypesTest.cpp
index f09ab5e124384..f04d45cf0983c 100644
--- a/llvm/unittests/Object/ELFTypesTest.cpp
+++ b/llvm/unittests/Object/ELFTypesTest.cpp
@@ -100,34 +100,36 @@ static_assert(
decltype(BBAddrMap::BBEntry::ID)>,
"PGOAnalysisMap should use the same type for basic block ID as BBAddrMap");
-TEST(ELFTypesTest, PGOAnalysisMapFeaturesEncodingTest) {
- const std::array<PGOAnalysisMap::Features, 7> Decoded = {
- {{false, false, false},
- {true, false, false},
- {false, true, false},
- {false, false, true},
- {true, true, false},
- {false, true, true},
- {true, true, true}}};
- const std::array<uint8_t, 7> Encoded = {
- {0b000, 0b001, 0b010, 0b100, 0b011, 0b110, 0b111}};
+TEST(ELFTypesTest, BBAddrMapFeaturesEncodingTest) {
+ const std::array<BBAddrMap::Features, 9> Decoded = {
+ {{false, false, false, false},
+ {true, false, false, false},
+ {false, true, false, false},
+ {false, false, true, false},
+ {false, false, false, true},
+ {true, true, false, false},
+ {false, true, true, false},
+ {false, true, true, true},
+ {true, true, true, true}}};
+ const std::array<uint8_t, 9> Encoded = {
+ {0b0000, 0b0001, 0b0010, 0b0100, 0b1000, 0b0011, 0b0110, 0b1110, 0b1111}};
for (const auto &[Feat, EncodedVal] : llvm::zip(Decoded, Encoded))
EXPECT_EQ(Feat.encode(), EncodedVal);
for (const auto &[Feat, EncodedVal] : llvm::zip(Decoded, Encoded)) {
- Expected<PGOAnalysisMap::Features> FeatEnableOrError =
- PGOAnalysisMap::Features::decode(EncodedVal);
+ Expected<BBAddrMap::Features> FeatEnableOrError =
+ BBAddrMap::Features::decode(EncodedVal);
ASSERT_THAT_EXPECTED(FeatEnableOrError, Succeeded());
EXPECT_EQ(*FeatEnableOrError, Feat);
}
}
-TEST(ELFTypesTest, PGOAnalysisMapFeaturesInvalidEncodingTest) {
+TEST(ELFTypesTest, BBAddrMapFeaturesInvalidEncodingTest) {
const std::array<std::string, 2> Errors = {
- "invalid encoding for PGOAnalysisMap::Features: 0x8",
- "invalid encoding for PGOAnalysisMap::Features: 0xff"};
- const std::array<uint8_t, 2> Values = {{0b1000, 0b1111'1111}};
+ "invalid encoding for BBAddrMap::Features: 0x10",
+ "invalid encoding for BBAddrMap::Features: 0xff"};
+ const std::array<uint8_t, 2> Values = {{0b10000, 0b1111'1111}};
for (const auto &[Val, Error] : llvm::zip(Values, Errors)) {
- EXPECT_THAT_ERROR(PGOAnalysisMap::Features::decode(Val).takeError(),
+ EXPECT_THAT_ERROR(BBAddrMap::Features::decode(Val).takeError(),
FailedWithMessage(Error));
}
}
>From 43dde6939bf557b719cd8a62bde39543756cfa73 Mon Sep 17 00:00:00 2001
From: Rahman Lavaee <rahmanl at google.com>
Date: Tue, 23 Jan 2024 18:50:51 +0000
Subject: [PATCH 10/12] Add obj2yaml test for empty BBRanges.
---
llvm/include/llvm/Object/ELFTypes.h | 7 +-
llvm/include/llvm/Target/TargetOptions.h | 7 ++
llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml | 2 +-
llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml | 17 +++
llvm/tools/llvm-objdump/llvm-objdump.cpp | 119 +++++++++---------
5 files changed, 91 insertions(+), 61 deletions(-)
diff --git a/llvm/include/llvm/Object/ELFTypes.h b/llvm/include/llvm/Object/ELFTypes.h
index 810f2156e5f5a..8a9a6ef9a9f6d 100644
--- a/llvm/include/llvm/Object/ELFTypes.h
+++ b/llvm/include/llvm/Object/ELFTypes.h
@@ -929,15 +929,16 @@ struct BBAddrMap {
// Returns the total number of bb entries in all bb ranges.
size_t getNumBBEntries() const {
size_t NumBBEntries = 0;
- for (const auto &BBR: BBRanges)
+ for (const auto &BBR : BBRanges)
NumBBEntries += BBR.BBEntries.size();
return NumBBEntries;
}
// Returns the index of the bb range with the given base address, or
// `std::nullopt` if no such range exists.
- std::optional<size_t> getBBRangeIndexForBaseAddress(uint64_t BaseAddress) const {
- for (size_t I=0; I<BBRanges.size(); ++I)
+ std::optional<size_t>
+ getBBRangeIndexForBaseAddress(uint64_t BaseAddress) const {
+ for (size_t I = 0; I < BBRanges.size(); ++I)
if (BBRanges[I].BaseAddress == BaseAddress)
return I;
return {};
diff --git a/llvm/include/llvm/Target/TargetOptions.h b/llvm/include/llvm/Target/TargetOptions.h
index 4bc03b0dc80a4..7df8010d55c70 100644
--- a/llvm/include/llvm/Target/TargetOptions.h
+++ b/llvm/include/llvm/Target/TargetOptions.h
@@ -295,6 +295,9 @@ namespace llvm {
/// function in the runtime library..
unsigned EmulatedTLS : 1;
+ /// EnableTLSDESC - This flag enables TLS Descriptors.
+ unsigned EnableTLSDESC : 1;
+
/// This flag enables InterProcedural Register Allocation (IPRA).
unsigned EnableIPRA : 1;
@@ -313,6 +316,10 @@ namespace llvm {
/// Emit address-significance table.
unsigned EmitAddrsig : 1;
+ // Emit the SHT_LLVM_BB_ADDR_MAP section containing basic block address
+ // which can be used to map virtual addresses to machine basic blocks.
+ unsigned BBAddrMap : 1;
+
/// Emit basic blocks into separate sections.
BasicBlockSection BBSections = BasicBlockSection::None;
diff --git a/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml b/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
index f9b3c30c158e4..8dbf97ef2bc12 100644
--- a/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
+++ b/llvm/test/tools/obj2yaml/ELF/bb-addr-map.yaml
@@ -162,7 +162,7 @@ Sections:
## Check that obj2yaml uses the "Content" tag to describe an .llvm_bb_addr_map section
## when it can't extract the entries, for example, when the section is truncated, or
-## when an invalid 'NumBlocks' field is specified.
+## when an invalid 'NumBlocks' or 'NumBBRanges` field is specified.
# RUN: yaml2obj --docnum=1 -DSIZE=0x8 %s -o %t4
# RUN: obj2yaml %t4 | FileCheck %s --check-prefixes=TRUNCATED,INVALID
diff --git a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
index 74357a4cfa605..27f9c487eb36d 100644
--- a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
+++ b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
@@ -53,6 +53,14 @@
# CHECK-NEXT: 0000: 02002000 00000000 0000020D 010203
# CHECK-NEXT: )
+# Case 7: Specify empty BBRanges.
+# CHECK: Name: .llvm_bb_addr_map (1)
+# CHECK: SectionData (
+# CHECK-NEXT: 0000: 0200
+# CHECK-NEXT: )
+
+
+
--- !ELF
FileHeader:
Class: ELFCLASS64
@@ -122,6 +130,15 @@ Sections:
Size: 0x00000002
Metadata: 0x00000003
+## 7) We can produce a SHT_LLVM_BB_ADDR_MAP section from a description with one entry with empty BBRanges.
+ - Name: '.llvm_bb_addr_map (7)'
+ Type: SHT_LLVM_BB_ADDR_MAP
+ Entries:
+ - Version: 2
+ BBRanges: []
+
+
+
## Check we can't use Entries at the same time as either Content or Size.
# RUN: not yaml2obj --docnum=2 -DCONTENT="00" %s 2>&1 | FileCheck %s --check-prefix=INVALID
# RUN: not yaml2obj --docnum=2 -DSIZE="0" %s 2>&1 | FileCheck %s --check-prefix=INVALID
diff --git a/llvm/tools/llvm-objdump/llvm-objdump.cpp b/llvm/tools/llvm-objdump/llvm-objdump.cpp
index d0b095c02f507..ffbc53af5b227 100644
--- a/llvm/tools/llvm-objdump/llvm-objdump.cpp
+++ b/llvm/tools/llvm-objdump/llvm-objdump.cpp
@@ -177,83 +177,81 @@ struct BBAddrMapLabel {
std::string PGOAnalysis;
};
-
// This class represents the BBAddrMap and PGOMap associated with a single
// function.
class BBAddrMapFunctionEntry {
- public:
- BBAddrMapFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap) : AddrMap(std::move(AddrMap)), PGOMap(std::move(PGOMap)) {}
+public:
+ BBAddrMapFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap)
+ : AddrMap(std::move(AddrMap)), PGOMap(std::move(PGOMap)) {}
- const BBAddrMap &getAddrMap() const {
- return AddrMap;
- }
+ const BBAddrMap &getAddrMap() const { return AddrMap; }
// Returns the PGO string associated with the entry of index `PGOBBEntryIndex`
// in `PGOMap`.
std::string constructPGOLabelString(size_t PGOBBEntryIndex) const {
if (!PGOMap.FeatEnable.hasPGOAnalysis())
return "";
- std::string PGOString;
- raw_string_ostream PGOSS(PGOString);
+ std::string PGOString;
+ raw_string_ostream PGOSS(PGOString);
- PGOSS << " (";
- if (PGOMap.FeatEnable.FuncEntryCount && PGOBBEntryIndex == 0) {
- PGOSS << "Entry count: " << Twine(PGOMap.FuncEntryCount);
- if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
- PGOSS << ", ";
+ PGOSS << " (";
+ if (PGOMap.FeatEnable.FuncEntryCount && PGOBBEntryIndex == 0) {
+ PGOSS << "Entry count: " << Twine(PGOMap.FuncEntryCount);
+ if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
+ PGOSS << ", ";
+ }
}
- }
- if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
+ if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
- assert(PGOBBEntryIndex < PGOMap.BBEntries.size() &&
- "Expected PGOAnalysisMap and BBAddrMap to have the same entires");
- const PGOAnalysisMap::PGOBBEntry &PGOBBEntry =
- PGOMap.BBEntries[PGOBBEntryIndex];
+ assert(PGOBBEntryIndex < PGOMap.BBEntries.size() &&
+ "Expected PGOAnalysisMap and BBAddrMap to have the same entires");
+ const PGOAnalysisMap::PGOBBEntry &PGOBBEntry =
+ PGOMap.BBEntries[PGOBBEntryIndex];
- if (PGOMap.FeatEnable.BBFreq) {
- PGOSS << "Frequency: " << Twine(PGOBBEntry.BlockFreq.getFrequency());
+ if (PGOMap.FeatEnable.BBFreq) {
+ PGOSS << "Frequency: " << Twine(PGOBBEntry.BlockFreq.getFrequency());
+ if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
+ PGOSS << ", ";
+ }
+ }
if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
- PGOSS << ", ";
+ PGOSS << "Successors: ";
+ interleaveComma(
+ PGOBBEntry.Successors, PGOSS,
+ [&PGOSS](const PGOAnalysisMap::PGOBBEntry::SuccessorEntry &SE) {
+ PGOSS << "BB" << SE.ID << ":";
+ PGOSS.write_hex(SE.Prob.getNumerator());
+ });
}
}
- if (PGOMap.FeatEnable.BrProb && PGOBBEntry.Successors.size() > 0) {
- PGOSS << "Successors: ";
- interleaveComma(
- PGOBBEntry.Successors, PGOSS,
- [&PGOSS](const PGOAnalysisMap::PGOBBEntry::SuccessorEntry &SE) {
- PGOSS << "BB" << SE.ID << ":";
- PGOSS.write_hex(SE.Prob.getNumerator());
- });
- }
+ PGOSS << ")";
+
+ return PGOString;
}
- PGOSS << ")";
- return PGOString;
-}
private:
- const BBAddrMap AddrMap;
- const PGOAnalysisMap PGOMap;
-
- };
+ const BBAddrMap AddrMap;
+ const PGOAnalysisMap PGOMap;
+};
// This class represents the BBAddrMap and PGOMap of potentially multiple
// functions in a section.
class BBAddrMapInfo {
- public:
+public:
void clear() {
FunctionAddrToMap.clear();
RangeBaseAddrToFunctionAddr.clear();
}
- bool empty() const {
- return FunctionAddrToMap.empty();
- }
+ bool empty() const { return FunctionAddrToMap.empty(); }
void AddFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap) {
uint64_t FunctionAddr = AddrMap.getFunctionAddress();
- for (size_t I=1; I < AddrMap.BBRanges.size(); ++I)
- RangeBaseAddrToFunctionAddr.emplace(AddrMap.BBRanges[I].BaseAddress, FunctionAddr);
- [[maybe_unused]] auto R = FunctionAddrToMap.try_emplace(FunctionAddr, std::move(AddrMap), std::move(PGOMap));
+ for (size_t I = 1; I < AddrMap.BBRanges.size(); ++I)
+ RangeBaseAddrToFunctionAddr.emplace(AddrMap.BBRanges[I].BaseAddress,
+ FunctionAddr);
+ [[maybe_unused]] auto R = FunctionAddrToMap.try_emplace(
+ FunctionAddr, std::move(AddrMap), std::move(PGOMap));
assert(R.second && "duplicate function address");
}
@@ -263,11 +261,12 @@ class BBAddrMapInfo {
if (S != RangeBaseAddrToFunctionAddr.end())
FunctionAddr = S->second;
auto R = FunctionAddrToMap.find(FunctionAddr);
- if (R == FunctionAddrToMap.end()) return nullptr;
+ if (R == FunctionAddrToMap.end())
+ return nullptr;
return &R->second;
}
- private:
+private:
std::unordered_map<uint64_t, BBAddrMapFunctionEntry> FunctionAddrToMap;
std::unordered_map<uint64_t, uint64_t> RangeBaseAddrToFunctionAddr;
};
@@ -1367,23 +1366,26 @@ static SymbolInfoTy createDummySymbolInfo(const ObjectFile &Obj,
}
static void collectBBAddrMapLabels(
- const BBAddrMapInfo &FullAddrMap,
- uint64_t SectionAddr, uint64_t Start, uint64_t End,
+ const BBAddrMapInfo &FullAddrMap, uint64_t SectionAddr, uint64_t Start,
+ uint64_t End,
std::unordered_map<uint64_t, std::vector<BBAddrMapLabel>> &Labels) {
if (FullAddrMap.empty())
return;
Labels.clear();
uint64_t StartAddress = SectionAddr + Start;
uint64_t EndAddress = SectionAddr + End;
- const BBAddrMapFunctionEntry *FunctionMap = FullAddrMap.getEntryForAddress(StartAddress);
+ const BBAddrMapFunctionEntry *FunctionMap =
+ FullAddrMap.getEntryForAddress(StartAddress);
if (!FunctionMap)
return;
- std::optional<size_t> BBRangeIndex = FunctionMap->getAddrMap().getBBRangeIndexForBaseAddress(StartAddress);
+ std::optional<size_t> BBRangeIndex =
+ FunctionMap->getAddrMap().getBBRangeIndexForBaseAddress(StartAddress);
if (!BBRangeIndex)
return;
size_t NumBBEntriesBeforeRange = 0;
- for (size_t I=0; I < *BBRangeIndex; ++I)
- NumBBEntriesBeforeRange += FunctionMap->getAddrMap().BBRanges[I].BBEntries.size();
+ for (size_t I = 0; I < *BBRangeIndex; ++I)
+ NumBBEntriesBeforeRange +=
+ FunctionMap->getAddrMap().BBRanges[I].BBEntries.size();
const auto &BBRange = FunctionMap->getAddrMap().BBRanges[*BBRangeIndex];
for (size_t I = 0; I < BBRange.BBEntries.size(); ++I) {
const BBAddrMap::BBEntry &BBEntry = BBRange.BBEntries[I];
@@ -1392,7 +1394,9 @@ static void collectBBAddrMapLabels(
continue;
std::string LabelString = ("BB" + Twine(BBEntry.ID)).str();
- Labels[BBAddress].push_back({LabelString, FunctionMap->constructPGOLabelString(NumBBEntriesBeforeRange + I)});
+ Labels[BBAddress].push_back(
+ {LabelString,
+ FunctionMap->constructPGOLabelString(NumBBEntriesBeforeRange + I)});
}
}
@@ -1762,7 +1766,8 @@ disassembleObject(ObjectFile &Obj, const ObjectFile &DbgObj,
}
for (auto &&[FunctionBBAddrMap, FunctionPGOAnalysis] :
zip_equal(*std::move(BBAddrMapsOrErr), std::move(PGOAnalyses))) {
- FullAddrMap.AddFunctionEntry(std::move(FunctionBBAddrMap), std::move(FunctionPGOAnalysis));
+ FullAddrMap.AddFunctionEntry(std::move(FunctionBBAddrMap),
+ std::move(FunctionPGOAnalysis));
}
}
};
@@ -2098,8 +2103,8 @@ disassembleObject(ObjectFile &Obj, const ObjectFile &DbgObj,
DT->DisAsm.get(), DT->InstPrinter.get(),
PrimaryTarget.SubtargetInfo.get(),
SectionAddr, Index, End, AllLabels);
- collectBBAddrMapLabels(FullAddrMap,
- SectionAddr, Index, End, BBAddrMapLabels);
+ collectBBAddrMapLabels(FullAddrMap, SectionAddr, Index, End,
+ BBAddrMapLabels);
}
if (DT->InstrAnalysis)
>From 28879e76a97e0a67cbff668589cfd2fd779c13b2 Mon Sep 17 00:00:00 2001
From: Rahman Lavaee <rahmanl at google.com>
Date: Thu, 25 Jan 2024 20:30:26 +0000
Subject: [PATCH 11/12] Add llvm-objdump tests for bbsections with bb-addr-map.
---
llvm/include/llvm/Object/ELFTypes.h | 29 ++--
llvm/lib/Object/ELF.cpp | 6 +
llvm/lib/ObjectYAML/ELFEmitter.cpp | 4 +-
...ddrmap-disassemble-symbolize-operands.yaml | 132 ++++++++++++------
.../elf-bbaddrmap-symbolize-relocatable.yaml | 32 ++++-
llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml | 21 ++-
llvm/tools/llvm-objdump/llvm-objdump.cpp | 7 +-
llvm/unittests/Object/ELFObjectFileTest.cpp | 13 ++
8 files changed, 180 insertions(+), 64 deletions(-)
diff --git a/llvm/include/llvm/Object/ELFTypes.h b/llvm/include/llvm/Object/ELFTypes.h
index 8a9a6ef9a9f6d..4986ecf8323d0 100644
--- a/llvm/include/llvm/Object/ELFTypes.h
+++ b/llvm/include/llvm/Object/ELFTypes.h
@@ -879,10 +879,11 @@ struct BBAddrMap {
}
};
- uint32_t ID; // Unique ID of this basic block.
- uint32_t Offset; // Offset of basic block relative to the base address.
- uint32_t Size; // Size of the basic block.
- Metadata MD; // Metdata for this basic block.
+ uint32_t ID = 0; // Unique ID of this basic block.
+ uint32_t Offset = 0; // Offset of basic block relative to the base address.
+ uint32_t Size = 0; // Size of the basic block.
+ Metadata MD = {false, false, false, false,
+ false}; // Metdata for this basic block.
BBEntry(uint32_t ID, uint32_t Offset, uint32_t Size, Metadata MD)
: ID(ID), Offset(Offset), Size(Size), MD(MD){};
@@ -902,7 +903,7 @@ struct BBAddrMap {
// Struct representing the BBAddrMap information for a contiguous range of
// basic blocks (a function or a basic block section).
struct BBRangeEntry {
- uint64_t BaseAddress; // Base address of the range.
+ uint64_t BaseAddress = 0; // Base address of the range.
std::vector<BBEntry> BBEntries; // Basic block entries for this range.
// Equality operator for unit testing.
@@ -913,16 +914,14 @@ struct BBAddrMap {
}
};
- // All ranges for this function. The first range always corresponds to the
- // function entry.
+ // All ranges for this function. Cannot be empty. The first range always
+ // corresponds to the function entry.
std::vector<BBRangeEntry> BBRanges;
// Returns the function address associated with this BBAddrMap, which is
- // stored as the `BaseAddress` of its first BBRangeEntry. Returns 0 if
- // BBRanges is empty.
+ // stored as the `BaseAddress` of its first BBRangeEntry.
uint64_t getFunctionAddress() const {
- if (BBRanges.empty())
- return 0;
+ assert(!BBRanges.empty());
return BBRanges.front().BaseAddress;
}
@@ -944,11 +943,13 @@ struct BBAddrMap {
return {};
}
- // Returns bb entries in the bb range indexed `BBRangeIndex`.
- const std::vector<BBEntry> &getBBEntries(unsigned BBRangeIndex = 0) const {
- return BBRanges.at(BBRangeIndex).BBEntries;
+ // Returns bb entries in the first range.
+ const std::vector<BBEntry> &getBBEntries() const {
+ return BBRanges.front().BBEntries;
}
+ const std::vector<BBRangeEntry> &getBBRanges() const { return BBRanges; }
+
// Equality operator for unit testing.
bool operator==(const BBAddrMap &Other) const {
return std::equal(BBRanges.begin(), BBRanges.end(), Other.BBRanges.begin());
diff --git a/llvm/lib/Object/ELF.cpp b/llvm/lib/Object/ELF.cpp
index d7f75a6303673..137f606dd2d46 100644
--- a/llvm/lib/Object/ELF.cpp
+++ b/llvm/lib/Object/ELF.cpp
@@ -764,6 +764,12 @@ decodeBBAddrMapImpl(const ELFFile<ELFT> &EF,
std::vector<BBAddrMap::BBEntry> BBEntries;
if (FeatEnable.MultiBBRange) {
NumBBRanges = readULEB128As<uint32_t>(Data, Cur, ULEBSizeErr);
+ if (!Cur || ULEBSizeErr)
+ break;
+ if (!NumBBRanges)
+ return createError("invalid zero number of BB ranges at offset " +
+ Twine::utohexstr(Cur.tell()) + " in " +
+ describe(EF, Sec));
} else {
auto AddressOrErr = ExtractAddress();
if (!AddressOrErr)
diff --git a/llvm/lib/ObjectYAML/ELFEmitter.cpp b/llvm/lib/ObjectYAML/ELFEmitter.cpp
index 3e32f8c89d841..58a725f8d8778 100644
--- a/llvm/lib/ObjectYAML/ELFEmitter.cpp
+++ b/llvm/lib/ObjectYAML/ELFEmitter.cpp
@@ -1427,8 +1427,8 @@ void ELFState<ELFT>::writeSectionContent(
MultiBBRangeFeatureEnabled = FeatureOrErr->MultiBBRange;
bool MultiBBRange =
MultiBBRangeFeatureEnabled ||
- (E.NumBBRanges.has_value() && E.NumBBRanges.value() > 1) ||
- (E.BBRanges && E.BBRanges->size() > 1);
+ (E.NumBBRanges.has_value() && E.NumBBRanges.value() != 1) ||
+ (E.BBRanges && E.BBRanges->size() != 1);
if (MultiBBRange && !MultiBBRangeFeatureEnabled)
WithColor::warning() << "feature value(" << E.Feature
<< ") does not support multiple BB ranges.";
diff --git a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
index cc977af76dfb4..cc7faea67bed2 100644
--- a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
+++ b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-disassemble-symbolize-operands.yaml
@@ -1,30 +1,32 @@
## Test that in the presence of SHT_LLVM_BB_ADDR_MAP sections,
## --symbolize-operands can display <BB*> labels.
-## Executable object file.
-# RUN: yaml2obj --docnum=1 -DFOO_ADDR=0x4000 -DBAR_ADDR=0x5000 %s -o %t1
-# RUN: llvm-objdump %t1 -d --symbolize-operands -M intel --no-show-raw-insn --no-leading-addr | \
-# RUN: FileCheck %s -DSYM=symbol --match-full-lines --check-prefixes=INTEL
-# RUN: llvm-objdump %t1 -d --symbolize-operands -M att --no-show-raw-insn --no-leading-addr | \
-# RUN: FileCheck %s -DSYM=symbol --match-full-lines --check-prefixes=ATT
+## Executable object file with separate SHT_LLVM_BB_ADDR_MAP sections for multiple text sections.
+## RUN: yaml2obj --docnum=1 %s -o %t1
+## RUN: llvm-objdump %t1 -d --symbolize-operands -M intel --no-show-raw-insn --no-leading-addr | \
+## RUN: FileCheck %s --match-full-lines --check-prefixes=INTEL
+## RUN: llvm-objdump %t1 -d --symbolize-operands -M att --no-show-raw-insn --no-leading-addr | \
+## RUN: FileCheck %s --match-full-lines --check-prefixes=ATT
## Executable object file with a single SHT_LLVM_BB_ADDR_MAP for multiple text sections.
-# RUN: yaml2obj --docnum=2 %s -o %t3
-# RUN: llvm-objdump %t3 -d --symbolize-operands -M intel --no-show-raw-insn --no-leading-addr | \
-# RUN: FileCheck %s -DSYM=symbol --match-full-lines --check-prefixes=INTEL
-# RUN: llvm-objdump %t3 -d --symbolize-operands -M att --no-show-raw-insn --no-leading-addr | \
-# RUN: FileCheck %s -DSYM=symbol --match-full-lines --check-prefixes=ATT
+# RUN: yaml2obj --docnum=2 %s -o %t2
+# RUN: llvm-objdump %t2 -d --symbolize-operands -M intel --no-show-raw-insn --no-leading-addr | \
+# RUN: FileCheck %s --match-full-lines --check-prefixes=INTEL
+# RUN: llvm-objdump %t2 -d --symbolize-operands -M att --no-show-raw-insn --no-leading-addr | \
+# RUN: FileCheck %s --match-full-lines --check-prefixes=ATT
## Expect to find the branch and basic block labels and global variable name.
# ATT: <foo>:
# ATT-NEXT: <BB3>:
# ATT-NEXT: pushq %rax
# ATT-NEXT: <BB1>:
-# ATT-NEXT: cmpl , %eax <[[SYM]]>
+# ATT-NEXT: cmpl , %eax <symbol>
# ATT-NEXT: nop
# ATT-NEXT: <BB2>:
# ATT-NEXT: jge <BB5>
# ATT-NEXT: jmp <BB1>
+# ATT-NEXT: <BB4>:
+# ATT-NEXT: jge <foo.cold>
# ATT-NEXT: <BB5>:
# ATT-NEXT: retq
# ATT: <bar>:
@@ -37,16 +39,21 @@
# ATT-NEXT: <BB2>:
# ATT-NEXT: callq <bar>
# ATT-NEXT: retq
+# ATT: <foo.cold>:
+# ATT-NEXT: <BB6>:
+# ATT-NEXT: retq
# INTEL: <foo>:
# INTEL-NEXT: <BB3>:
# INTEL-NEXT: push rax
# INTEL-NEXT: <BB1>:
-# INTEL-NEXT: cmp eax, dword ptr <[[SYM]]>
+# INTEL-NEXT: cmp eax, dword ptr <symbol>
# INTEL-NEXT: nop
# INTEL-NEXT: <BB2>:
# INTEL-NEXT: jge <BB5>
# INTEL-NEXT: jmp <BB1>
+# INTEL-NEXT: <BB4>:
+# INTEL-NEXT: jge <foo.cold>
# INTEL-NEXT: <BB5>:
# INTEL-NEXT: ret
# INTEL: <bar>:
@@ -59,6 +66,9 @@
# INTEL-NEXT: <BB2>:
# INTEL-NEXT: call <bar>
# INTEL-NEXT: ret
+# INTEL: <foo.cold>:
+# INTEL-NEXT: <BB6>:
+# INTEL-NEXT: ret
## This object file contains a separate text section and SHT_LLVM_BB_ADDR_MAP
## section for each of the two functions foo and bar.
@@ -73,49 +83,65 @@ FileHeader:
Sections:
- Name: .text.foo
Type: SHT_PROGBITS
- Address: [[FOO_ADDR]]
+ Address: 0x4000
Flags: [SHF_ALLOC, SHF_EXECINSTR]
- Content: '503b0505200000907d02ebf5c3'
+ Content: '503b0505300000907d08ebf50f8dee1f0000c3'
- Name: .text.bar
Type: SHT_PROGBITS
- Address: [[BAR_ADDR]]
+ Address: 0x5000
Flags: [SHF_ALLOC, SHF_EXECINSTR]
Content: '5089d0740231f6e8f4ffffffc3'
+ - Name: .text.split
+ Type: SHT_PROGBITS
+ Address: 0x6000
+ Flags: [SHF_ALLOC, SHF_EXECINSTR]
+ Content: 'c3'
- Name: .data
Type: SHT_PROGBITS
Flags: [SHF_ALLOC, SHF_WRITE]
- Address: 0x6000
+ Address: 0x7000
- Name: .llvm_bb_addr_map.foo
Type: SHT_LLVM_BB_ADDR_MAP
Link: .text.foo
Entries:
- Version: 2
+ Feature: 0x8
BBRanges:
- - BaseAddress: [[FOO_ADDR]]
- BBEntries:
- - ID: 3
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x1
- - ID: 1
- AddressOffset: 0x0
- Size: 0x6
- Metadata: 0x0
- - ID: 2
- AddressOffset: 0x1
- Size: 0x4
- Metadata: 0x0
- - ID: 5
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
+ - BaseAddress: 0x4000
+ BBEntries:
+ - ID: 3
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x1
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x0
+ - ID: 2
+ AddressOffset: 0x1
+ Size: 0x4
+ Metadata: 0x0
+ - ID: 4
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x1
+ - ID: 5
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
+ - BaseAddress: 0x6000
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
- Name: .llvm_bb_addr_map.bar
Type: SHT_LLVM_BB_ADDR_MAP
Link: .text.bar
Entries:
- Version: 1
BBRanges:
- - BaseAddress: [[BAR_ADDR]]
+ - BaseAddress: 0x5000
BBEntries:
- AddressOffset: 0x0
Size: 0x1
@@ -130,13 +156,16 @@ Sections:
Symbols:
- Name: foo
Section: .text.foo
- Value: [[FOO_ADDR]]
+ Value: 0x4000
+ - Name: foo.cold
+ Section: .text.split
+ Value: 0x6000
- Name: bar
Section: .text.bar
- Value: [[BAR_ADDR]]
+ Value: 0x5000
- Name: symbol
Section: .data
- Value: 0x600c
+ Value: 0x700c
## This object file contains a single SHT_LLVM_BB_ADDR_MAP for two text
## sections .text.foo and .text.bar.
@@ -151,21 +180,27 @@ Sections:
Type: SHT_PROGBITS
Address: 0x4000
Flags: [SHF_ALLOC, SHF_EXECINSTR]
- Content: '503b0505200000907d02ebf5c3'
+ Content: '503b0505300000907d08ebf50f8dee1f0000c3'
- Name: .text.bar
Type: SHT_PROGBITS
Address: 0x5000
Flags: [SHF_ALLOC, SHF_EXECINSTR]
Content: '5089d0740231f6e8f4ffffffc3'
+ - Name: .text.split
+ Type: SHT_PROGBITS
+ Address: 0x6000
+ Flags: [SHF_ALLOC, SHF_EXECINSTR]
+ Content: 'c3'
- Name: .data
Type: SHT_PROGBITS
Flags: [SHF_ALLOC, SHF_WRITE]
- Address: 0x6000
+ Address: 0x7000
- Name: .llvm_bb_addr_map.foo
Type: SHT_LLVM_BB_ADDR_MAP
Link: .text.foo
Entries:
- Version: 2
+ Feature: 0x8
BBRanges:
- BaseAddress: 0x4000
BBEntries:
@@ -181,10 +216,20 @@ Sections:
AddressOffset: 0x1
Size: 0x4
Metadata: 0x0
+ - ID: 4
+ AddressOffset: 0x0
+ Size: 0x6
+ Metadata: 0x1
- ID: 5
AddressOffset: 0x0
Size: 0x1
Metadata: 0x2
+ - BaseAddress: 0x6000
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
- Version: 1
BBRanges:
- BaseAddress: 0x5000
@@ -203,9 +248,12 @@ Symbols:
- Name: foo
Section: .text.foo
Value: 0x4000
+ - Name: foo.cold
+ Section: .text.split
+ Value: 0x6000
- Name: bar
Section: .text.bar
Value: 0x5000
- Name: symbol
Section: .data
- Value: 0x600c
+ Value: 0x700c
diff --git a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
index 88afaceac8c51..706d386e467e7 100644
--- a/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
+++ b/llvm/test/tools/llvm-objdump/X86/elf-bbaddrmap-symbolize-relocatable.yaml
@@ -29,14 +29,27 @@ Sections:
- ID: 0
AddressOffset: 0x0
Size: 0x1
- Metadata: 0x1
+ Metadata: 0xa
- Version: 2
BBRanges:
- BBEntries:
- ID: 0
AddressOffset: 0x0
Size: 0x1
- Metadata: 0x1
+ Metadata: 0xb
+ - Version: 2
+ Feature: 0x8
+ BBRanges:
+ - BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0xc
+ - BBEntries:
+ - ID: 1
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0xd
- Name: .rela.llvm_bb_addr_map
Type: SHT_RELA
Flags: [ SHF_INFO_LINK ]
@@ -50,6 +63,15 @@ Sections:
Symbol: .text
Type: R_X86_64_64
Addend: 1
+ - Offset: 0x21
+ Symbol: .text
+ Type: R_X86_64_64
+ Addend: 0x2
+ - Offset: 0x2e
+ Symbol: .text
+ Type: R_X86_64_64
+ Addend: 0x3
+
Symbols:
- Name: a
Section: .text
@@ -57,6 +79,12 @@ Symbols:
- Name: c
Section: .text
Value: 0x1
+ - Name: h
+ Section: .text
+ Value: 0x2
+ - Name: h1
+ Section: .text
+ Value: 0x3
- Name: .text
Type: STT_SECTION
Section: .text
diff --git a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
index 27f9c487eb36d..709938babffbf 100644
--- a/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
+++ b/llvm/test/tools/yaml2obj/ELF/bb-addr-map.yaml
@@ -56,9 +56,16 @@
# Case 7: Specify empty BBRanges.
# CHECK: Name: .llvm_bb_addr_map (1)
# CHECK: SectionData (
-# CHECK-NEXT: 0000: 0200
+# CHECK-NEXT: 0000: 020000
# CHECK-NEXT: )
+# Case 8: Specify empty BBRanges with multi-bb-range.
+# CHECK: Name: .llvm_bb_addr_map (1)
+# CHECK: SectionData (
+# CHECK-NEXT: 0000: 020800
+# CHECK-NEXT: )
+
+
--- !ELF
@@ -130,14 +137,22 @@ Sections:
Size: 0x00000002
Metadata: 0x00000003
-## 7) We can produce a SHT_LLVM_BB_ADDR_MAP section from a description with one entry with empty BBRanges.
+## 7) We can produce a SHT_LLVM_BB_ADDR_MAP section from a description
+## with one entry with empty BBRanges.
- Name: '.llvm_bb_addr_map (7)'
Type: SHT_LLVM_BB_ADDR_MAP
Entries:
- Version: 2
BBRanges: []
-
+## 8) We can produce a SHT_LLVM_BB_ADDR_MAP section from a multi-bb-range
+## description with one entry with empty BBRanges.
+ - Name: '.llvm_bb_addr_map (8)'
+ Type: SHT_LLVM_BB_ADDR_MAP
+ Entries:
+ - Version: 2
+ Feature: 0x8
+ BBRanges: []
## Check we can't use Entries at the same time as either Content or Size.
# RUN: not yaml2obj --docnum=2 -DCONTENT="00" %s 2>&1 | FileCheck %s --check-prefix=INVALID
diff --git a/llvm/tools/llvm-objdump/llvm-objdump.cpp b/llvm/tools/llvm-objdump/llvm-objdump.cpp
index ffbc53af5b227..b4467ec163c34 100644
--- a/llvm/tools/llvm-objdump/llvm-objdump.cpp
+++ b/llvm/tools/llvm-objdump/llvm-objdump.cpp
@@ -205,7 +205,7 @@ class BBAddrMapFunctionEntry {
if (PGOMap.FeatEnable.hasPGOAnalysisBBData()) {
assert(PGOBBEntryIndex < PGOMap.BBEntries.size() &&
- "Expected PGOAnalysisMap and BBAddrMap to have the same entires");
+ "Expected PGOAnalysisMap and BBAddrMap to have the same entries");
const PGOAnalysisMap::PGOBBEntry &PGOBBEntry =
PGOMap.BBEntries[PGOBBEntryIndex];
@@ -243,6 +243,7 @@ class BBAddrMapInfo {
FunctionAddrToMap.clear();
RangeBaseAddrToFunctionAddr.clear();
}
+
bool empty() const { return FunctionAddrToMap.empty(); }
void AddFunctionEntry(BBAddrMap AddrMap, PGOAnalysisMap PGOMap) {
@@ -255,6 +256,10 @@ class BBAddrMapInfo {
assert(R.second && "duplicate function address");
}
+ // Returns the BBAddrMap entry for the function associated with `BaseAddress`.
+ // `BaseAddress` could be the function address or the address of a range
+ // associated with that function. Returns `nullptr` if `BaseAddress` is not
+ // mapped to any entry.
const BBAddrMapFunctionEntry *getEntryForAddress(uint64_t BaseAddress) const {
uint64_t FunctionAddr = BaseAddress;
auto S = RangeBaseAddrToFunctionAddr.find(BaseAddress);
diff --git a/llvm/unittests/Object/ELFObjectFileTest.cpp b/llvm/unittests/Object/ELFObjectFileTest.cpp
index 99817913ca1be..c4d2b4ae8b9ac 100644
--- a/llvm/unittests/Object/ELFObjectFileTest.cpp
+++ b/llvm/unittests/Object/ELFObjectFileTest.cpp
@@ -545,6 +545,19 @@ TEST(ELFObjectFileTest, InvalidDecodeBBAddrMap) {
"unsupported SHT_LLVM_BB_ADDR_MAP version: 3");
}
+ SmallString<128> ZeroBBRangesYamlString(CommonYamlString);
+ ZeroBBRangesYamlString += R"(
+ - Version: 2
+ Feature: 0x8
+ BBRanges: []
+)";
+ {
+ SCOPED_TRACE("zero bb ranges");
+ DoCheck(ZeroBBRangesYamlString,
+ "invalid zero number of BB ranges at offset 3 in "
+ "SHT_LLVM_BB_ADDR_MAP section with index 1");
+ }
+
SmallString<128> CommonVersionedYamlString(CommonYamlString);
CommonVersionedYamlString += R"(
- Version: 2
>From 4fbbbbc6a605c2e1939b7eeb7aa48898dbe6184f Mon Sep 17 00:00:00 2001
From: Rahman Lavaee <rahmanl at google.com>
Date: Thu, 1 Feb 2024 21:58:07 +0000
Subject: [PATCH 12/12] Resolve llvm-readobj merge conflicts.
---
.../ELF/bb-addr-map-pgo-analysis-map.test | 164 ++++++++++--------
llvm/tools/llvm-readobj/ELFDumper.cpp | 46 ++---
2 files changed, 116 insertions(+), 94 deletions(-)
diff --git a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
index 91b3d7e3902e3..e5a9400c670c0 100644
--- a/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
+++ b/llvm/test/tools/llvm-readobj/ELF/bb-addr-map-pgo-analysis-map.test
@@ -24,26 +24,31 @@
# CHECK-NEXT: At: [[ADDR]]
# CHECK-NEXT: warning: '[[FILE]]': could not identify function symbol for address ([[ADDR]]) in SHT_LLVM_BB_ADDR_MAP section with index 3
# CHECK-NEXT: Name: <?>
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 0
-# CHECK-NEXT: Offset: 0x0
-# CHECK-NEXT: Size: 0x1
-# CHECK-NEXT: HasReturn: No
-# CHECK-NEXT: HasTailCall: Yes
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: No
-# CHECK-NEXT: HasIndirectBranch: No
-# CHECK-NEXT: }
-# CHECK-NEXT: {
-# CHECK-NEXT: ID: 2
-# CHECK-NEXT: Offset: 0x4
-# CHECK-NEXT: Size: 0x4
-# CHECK-NEXT: HasReturn: Yes
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: Yes
-# CHECK-NEXT: CanFallThrough: No
-# CHECK-NEXT: HasIndirectBranch: Yes
+# CHECK-NEXT: Base Address: [[ADDR]]
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 0
+# CHECK-NEXT: Offset: 0x0
+# CHECK-NEXT: Size: 0x1
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: Yes
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 2
+# CHECK-NEXT: Offset: 0x4
+# CHECK-NEXT: Size: 0x4
+# CHECK-NEXT: HasReturn: Yes
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: Yes
+# CHECK-NEXT: CanFallThrough: No
+# CHECK-NEXT: HasIndirectBranch: Yes
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: PGO analyses {
@@ -69,16 +74,21 @@
# CHECK-NEXT: Function {
# CHECK-NEXT: At: 0x22222
# CHECK-NEXT: Name: foo
-# CHECK-NEXT: BB entries [
+# CHECK-NEXT: BB Ranges [
# CHECK-NEXT: {
-# CHECK-NEXT: ID: 4
-# CHECK-NEXT: Offset: 0x6
-# CHECK-NEXT: Size: 0x7
-# CHECK-NEXT: HasReturn: No
-# CHECK-NEXT: HasTailCall: No
-# CHECK-NEXT: IsEHPad: No
-# CHECK-NEXT: CanFallThrough: Yes
-# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: Base Address: 0x22222
+# CHECK-NEXT: BB Entries [
+# CHECK-NEXT: {
+# CHECK-NEXT: ID: 4
+# CHECK-NEXT: Offset: 0x6
+# CHECK-NEXT: Size: 0x7
+# CHECK-NEXT: HasReturn: No
+# CHECK-NEXT: HasTailCall: No
+# CHECK-NEXT: IsEHPad: No
+# CHECK-NEXT: CanFallThrough: Yes
+# CHECK-NEXT: HasIndirectBranch: No
+# CHECK-NEXT: }
+# CHECK-NEXT: ]
# CHECK-NEXT: }
# CHECK-NEXT: ]
# CHECK-NEXT: PGO analyses {
@@ -102,26 +112,31 @@
# TRUNCATED-NEXT: Function {
# TRUNCATED-NEXT: At: 0x33333
# TRUNCATED-NEXT: Name: bar
-# TRUNCATED-NEXT: BB entries [
-# TRUNCATED-NEXT: {
-# TRUNCATED-NEXT: ID: 6
-# TRUNCATED-NEXT: Offset: 0x9
-# TRUNCATED-NEXT: Size: 0xA
-# TRUNCATED-NEXT: HasReturn: Yes
-# TRUNCATED-NEXT: HasTailCall: Yes
-# TRUNCATED-NEXT: IsEHPad: No
-# TRUNCATED-NEXT: CanFallThrough: Yes
-# TRUNCATED-NEXT: HasIndirectBranch: Yes
-# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: BB Ranges [
# TRUNCATED-NEXT: {
-# TRUNCATED-NEXT: ID: 7
-# TRUNCATED-NEXT: Offset: 0x1F
-# TRUNCATED-NEXT: Size: 0xD
-# TRUNCATED-NEXT: HasReturn: No
-# TRUNCATED-NEXT: HasTailCall: Yes
-# TRUNCATED-NEXT: IsEHPad: Yes
-# TRUNCATED-NEXT: CanFallThrough: Yes
-# TRUNCATED-NEXT: HasIndirectBranch: No
+# TRUNCATED-NEXT: Base Address: 0x33333
+# TRUNCATED-NEXT: BB Entries [
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 6
+# TRUNCATED-NEXT: Offset: 0x9
+# TRUNCATED-NEXT: Size: 0xA
+# TRUNCATED-NEXT: HasReturn: Yes
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: No
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: Yes
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: {
+# TRUNCATED-NEXT: ID: 7
+# TRUNCATED-NEXT: Offset: 0x1F
+# TRUNCATED-NEXT: Size: 0xD
+# TRUNCATED-NEXT: HasReturn: No
+# TRUNCATED-NEXT: HasTailCall: Yes
+# TRUNCATED-NEXT: IsEHPad: Yes
+# TRUNCATED-NEXT: CanFallThrough: Yes
+# TRUNCATED-NEXT: HasIndirectBranch: No
+# TRUNCATED-NEXT: }
+# TRUNCATED-NEXT: ]
# TRUNCATED-NEXT: }
# TRUNCATED-NEXT: ]
# TRUNCATED-NEXT: PGO analyses {
@@ -151,24 +166,26 @@ Sections:
Entries:
- Version: 2
Feature: 0x7
- Address: [[ADDR=0x11111]]
- BBEntries:
- - ID: 0
- AddressOffset: 0x0
- Size: 0x1
- Metadata: 0x2
- - ID: 2
- AddressOffset: 0x3
- Size: 0x4
- Metadata: 0x15
+ BBRanges:
+ - BaseAddress: [[ADDR=0x11111]]
+ BBEntries:
+ - ID: 0
+ AddressOffset: 0x0
+ Size: 0x1
+ Metadata: 0x2
+ - ID: 2
+ AddressOffset: 0x3
+ Size: 0x4
+ Metadata: 0x15
- Version: 2
Feature: 0x3
- Address: 0x22222
- BBEntries:
- - ID: 4
- AddressOffset: 0x6
- Size: 0x7
- Metadata: 0x8
+ BBRanges:
+ - BaseAddress: 0x22222
+ BBEntries:
+ - ID: 4
+ AddressOffset: 0x6
+ Size: 0x7
+ Metadata: 0x8
PGOAnalyses:
- FuncEntryCount: 100
PGOBBEntries:
@@ -190,16 +207,17 @@ Sections:
Entries:
- Version: 2
Feature: [[FEATURE=0x1]]
- Address: 0x33333
- BBEntries:
- - ID: 6
- AddressOffset: 0x9
- Size: 0xa
- Metadata: 0x1b
- - ID: 7
- AddressOffset: 0xc
- Size: 0xd
- Metadata: 0xe
+ BBRanges:
+ - BaseAddress: 0x33333
+ BBEntries:
+ - ID: 6
+ AddressOffset: 0x9
+ Size: 0xa
+ Metadata: 0x1b
+ - ID: 7
+ AddressOffset: 0xc
+ Size: 0xd
+ Metadata: 0xe
PGOAnalyses:
- FuncEntryCount: 89
Symbols:
diff --git a/llvm/tools/llvm-readobj/ELFDumper.cpp b/llvm/tools/llvm-readobj/ELFDumper.cpp
index 1005c81ff8fdf..ce33b15b099aa 100644
--- a/llvm/tools/llvm-readobj/ELFDumper.cpp
+++ b/llvm/tools/llvm-readobj/ELFDumper.cpp
@@ -7574,44 +7574,48 @@ template <class ELFT> void LLVMELFDumper<ELFT>::printBBAddrMaps() {
": " + toString(BBAddrMapOrErr.takeError()));
continue;
}
- for (const BBAddrMap &AM : *BBAddrMapOrErr) {
+ for (const auto &[AM, PAM] : zip_equal(*BBAddrMapOrErr, PGOAnalyses)) {
DictScope D(W, "Function");
- W.printHex("At", AM.Addr);
+ W.printHex("At", AM.getFunctionAddress());
SmallVector<uint32_t> FuncSymIndex =
- this->getSymbolIndexesForFunctionAddress(AM.Addr, FunctionSec);
+ this->getSymbolIndexesForFunctionAddress(AM.getFunctionAddress(),
+ FunctionSec);
std::string FuncName = "<?>";
if (FuncSymIndex.empty())
this->reportUniqueWarning(
"could not identify function symbol for address (0x" +
- Twine::utohexstr(AM.Addr) + ") in " + this->describe(*Sec));
+ Twine::utohexstr(AM.getFunctionAddress()) + ") in " +
+ this->describe(*Sec));
else
FuncName = this->getStaticSymbolName(FuncSymIndex.front());
W.printString("Name", FuncName);
- ListScope BBRL(W, "BB Ranges");
- for (const BBAddrMap::BBRangeEntry &BBR : AM.BBRanges) {
- DictScope BBRD(W);
- W.printHex("Base Address", BBR.BaseAddress);
- ListScope BBEL(W, "BB Entries");
- for (const BBAddrMap::BBEntry &BBE : BBR.BBEntries) {
- DictScope BBED(W);
- W.printNumber("ID", BBE.ID);
- W.printHex("Offset", BBE.Offset);
- W.printHex("Size", BBE.Size);
- W.printBoolean("HasReturn", BBE.hasReturn());
- W.printBoolean("HasTailCall", BBE.hasTailCall());
- W.printBoolean("IsEHPad", BBE.isEHPad());
- W.printBoolean("CanFallThrough", BBE.canFallThrough());
- W.printBoolean("HasIndirectBranch", BBE.hasIndirectBranch());
+ {
+ ListScope BBRL(W, "BB Ranges");
+ for (const BBAddrMap::BBRangeEntry &BBR : AM.BBRanges) {
+ DictScope BBRD(W);
+ W.printHex("Base Address", BBR.BaseAddress);
+ ListScope BBEL(W, "BB Entries");
+ for (const BBAddrMap::BBEntry &BBE : BBR.BBEntries) {
+ DictScope BBED(W);
+ W.printNumber("ID", BBE.ID);
+ W.printHex("Offset", BBE.Offset);
+ W.printHex("Size", BBE.Size);
+ W.printBoolean("HasReturn", BBE.hasReturn());
+ W.printBoolean("HasTailCall", BBE.hasTailCall());
+ W.printBoolean("IsEHPad", BBE.isEHPad());
+ W.printBoolean("CanFallThrough", BBE.canFallThrough());
+ W.printBoolean("HasIndirectBranch", BBE.hasIndirectBranch());
+ }
}
}
- if (PAM.FeatEnable.anyEnabled()) {
+ if (PAM.FeatEnable.hasPGOAnalysis()) {
DictScope PD(W, "PGO analyses");
if (PAM.FeatEnable.FuncEntryCount)
W.printNumber("FuncEntryCount", PAM.FuncEntryCount);
- if (PAM.FeatEnable.BBFreq || PAM.FeatEnable.BrProb) {
+ if (PAM.FeatEnable.hasPGOAnalysisBBData()) {
ListScope L(W, "PGO BB entries");
for (const PGOAnalysisMap::PGOBBEntry &PBBE : PAM.BBEntries) {
DictScope L(W);
More information about the cfe-commits
mailing list