Merge branch 'akpm' (patches from Andrew)

Merge updates from Andrew Morton:

 - misc fixes

 - ocfs2 updates

 - most of MM

* emailed patches from Andrew Morton <akpm@linux-foundation.org>: (118 commits)
  mm: remove PG_highmem description
  tools, vm: new option to specify kpageflags file
  mm/swap.c: make functions and their kernel-doc agree
  mm, memory_hotplug: fix memmap initialization
  mm: correct comments regarding do_fault_around()
  mm: numa: do not trap faults on shared data section pages.
  hugetlb, mbind: fall back to default policy if vma is NULL
  hugetlb, mempolicy: fix the mbind hugetlb migration
  mm, hugetlb: further simplify hugetlb allocation API
  mm, hugetlb: get rid of surplus page accounting tricks
  mm, hugetlb: do not rely on overcommit limit during migration
  mm, hugetlb: integrate giga hugetlb more naturally to the allocation path
  mm, hugetlb: unify core page allocation accounting and initialization
  mm/memcontrol.c: try harder to decrease [memory,memsw].limit_in_bytes
  mm/memcontrol.c: make local symbol static
  mm/hmm: fix uninitialized use of 'entry' in hmm_vma_walk_pmd()
  include/linux/mmzone.h: fix explanation of lower bits in the SPARSEMEM mem_map pointer
  mm/compaction.c: fix comment for try_to_compact_pages()
  mm/page_ext.c: make page_ext_init a noop when CONFIG_PAGE_EXTENSION but nothing uses it
  zsmalloc: use U suffix for negative literals being shifted
  ...

12 files changed, 458 insertions, 379 deletions

diff --git a/tools/testing/selftests/memfd/Makefile b/tools/testing/selftests/memfd/Makefile
index 3926a0409dda..a5276a91dfbf 100644
--- a/tools/testing/selftests/memfd/Makefile
+++ b/tools/testing/selftests/memfd/Makefile
@@ -12,3 +12,8 @@ fuse_mnt.o: CFLAGS += $(shell pkg-config fuse --cflags)
 include ../lib.mk
 
 $(OUTPUT)/fuse_mnt: LDLIBS += $(shell pkg-config fuse --libs)
+
+$(OUTPUT)/memfd_test: memfd_test.c common.o
+$(OUTPUT)/fuse_test: fuse_test.c common.o
+
+EXTRA_CLEAN = common.o
diff --git a/tools/testing/selftests/memfd/common.c b/tools/testing/selftests/memfd/common.c
new file mode 100644
index 000000000000..8eb3d75f6e60
--- /dev/null
+++ b/tools/testing/selftests/memfd/common.c
@@ -0,0 +1,46 @@
+// SPDX-License-Identifier: GPL-2.0
+#define _GNU_SOURCE
+#define __EXPORTED_HEADERS__
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <linux/fcntl.h>
+#include <linux/memfd.h>
+#include <unistd.h>
+#include <sys/syscall.h>
+
+#include "common.h"
+
+int hugetlbfs_test = 0;
+
+/*
+ * Copied from mlock2-tests.c
+ */
+unsigned long default_huge_page_size(void)
+{
+	unsigned long hps = 0;
+	char *line = NULL;
+	size_t linelen = 0;
+	FILE *f = fopen("/proc/meminfo", "r");
+
+	if (!f)
+		return 0;
+	while (getline(&line, &linelen, f) > 0) {
+		if (sscanf(line, "Hugepagesize:       %lu kB", &hps) == 1) {
+			hps <<= 10;
+			break;
+		}
+	}
+
+	free(line);
+	fclose(f);
+	return hps;
+}
+
+int sys_memfd_create(const char *name, unsigned int flags)
+{
+	if (hugetlbfs_test)
+		flags |= MFD_HUGETLB;
+
+	return syscall(__NR_memfd_create, name, flags);
+}
diff --git a/tools/testing/selftests/memfd/common.h b/tools/testing/selftests/memfd/common.h
new file mode 100644
index 000000000000..522d2c630bd8
--- /dev/null
+++ b/tools/testing/selftests/memfd/common.h
@@ -0,0 +1,9 @@
+#ifndef COMMON_H_
+#define COMMON_H_
+
+extern int hugetlbfs_test;
+
+unsigned long default_huge_page_size(void);
+int sys_memfd_create(const char *name, unsigned int flags);
+
+#endif
diff --git a/tools/testing/selftests/memfd/fuse_test.c b/tools/testing/selftests/memfd/fuse_test.c
index 1ccb7a3eb14b..b018e835737d 100644
--- a/tools/testing/selftests/memfd/fuse_test.c
+++ b/tools/testing/selftests/memfd/fuse_test.c
@@ -33,14 +33,12 @@
 #include <sys/wait.h>
 #include <unistd.h>
 
+#include "common.h"
+
 #define MFD_DEF_SIZE 8192
 #define STACK_SIZE 65536
 
-static int sys_memfd_create(const char *name,
-			    unsigned int flags)
-{
-	return syscall(__NR_memfd_create, name, flags);
-}
+static size_t mfd_def_size = MFD_DEF_SIZE;
 
 static int mfd_assert_new(const char *name, loff_t sz, unsigned int flags)
 {
@@ -127,7 +125,7 @@ static void *mfd_assert_mmap_shared(int fd)
 	void *p;
 
 	p = mmap(NULL,
-		 MFD_DEF_SIZE,
+		 mfd_def_size,
 		 PROT_READ | PROT_WRITE,
 		 MAP_SHARED,
 		 fd,
@@ -145,7 +143,7 @@ static void *mfd_assert_mmap_private(int fd)
 	void *p;
 
 	p = mmap(NULL,
-		 MFD_DEF_SIZE,
+		 mfd_def_size,
 		 PROT_READ | PROT_WRITE,
 		 MAP_PRIVATE,
 		 fd,
@@ -178,7 +176,7 @@ static int sealing_thread_fn(void *arg)
 	usleep(200000);
 
 	/* unmount mapping before sealing to avoid i_mmap_writable failures */
-	munmap(global_p, MFD_DEF_SIZE);
+	munmap(global_p, mfd_def_size);
 
 	/* Try sealing the global file; expect EBUSY or success. Current
 	 * kernels will never succeed, but in the future, kernels might
@@ -228,7 +226,7 @@ static void join_sealing_thread(pid_t pid)
 
 int main(int argc, char **argv)
 {
-	static const char zero[MFD_DEF_SIZE];
+	char *zero;
 	int fd, mfd, r;
 	void *p;
 	int was_sealed;
@@ -239,6 +237,25 @@ int main(int argc, char **argv)
 		abort();
 	}
 
+	if (argc >= 3) {
+		if (!strcmp(argv[2], "hugetlbfs")) {
+			unsigned long hpage_size = default_huge_page_size();
+
+			if (!hpage_size) {
+				printf("Unable to determine huge page size\n");
+				abort();
+			}
+
+			hugetlbfs_test = 1;
+			mfd_def_size = hpage_size * 2;
+		} else {
+			printf("Unknown option: %s\n", argv[2]);
+			abort();
+		}
+	}
+
+	zero = calloc(sizeof(*zero), mfd_def_size);
+
 	/* open FUSE memfd file for GUP testing */
 	printf("opening: %s\n", argv[1]);
 	fd = open(argv[1], O_RDONLY | O_CLOEXEC);
@@ -249,7 +266,7 @@ int main(int argc, char **argv)
 
 	/* create new memfd-object */
 	mfd = mfd_assert_new("kern_memfd_fuse",
-			     MFD_DEF_SIZE,
+			     mfd_def_size,
 			     MFD_CLOEXEC | MFD_ALLOW_SEALING);
 
 	/* mmap memfd-object for writing */
@@ -268,7 +285,7 @@ int main(int argc, char **argv)
 	 * This guarantees that the receive-buffer is pinned for 1s until the
 	 * data is written into it. The racing ADD_SEALS should thus fail as
 	 * the pages are still pinned. */
-	r = read(fd, p, MFD_DEF_SIZE);
+	r = read(fd, p, mfd_def_size);
 	if (r < 0) {
 		printf("read() failed: %m\n");
 		abort();
@@ -295,10 +312,10 @@ int main(int argc, char **argv)
 	 * enough to avoid any in-flight writes. */
 
 	p = mfd_assert_mmap_private(mfd);
-	if (was_sealed && memcmp(p, zero, MFD_DEF_SIZE)) {
+	if (was_sealed && memcmp(p, zero, mfd_def_size)) {
 		printf("memfd sealed during read() but data not discarded\n");
 		abort();
-	} else if (!was_sealed && !memcmp(p, zero, MFD_DEF_SIZE)) {
+	} else if (!was_sealed && !memcmp(p, zero, mfd_def_size)) {
 		printf("memfd sealed after read() but data discarded\n");
 		abort();
 	}
@@ -307,6 +324,7 @@ int main(int argc, char **argv)
 	close(fd);
 
 	printf("fuse: DONE\n");
+	free(zero);
 
 	return 0;
 }
diff --git a/tools/testing/selftests/memfd/memfd_test.c b/tools/testing/selftests/memfd/memfd_test.c
index 132a54f74e88..10baa1652fc2 100644
--- a/tools/testing/selftests/memfd/memfd_test.c
+++ b/tools/testing/selftests/memfd/memfd_test.c
@@ -19,7 +19,10 @@
 #include <sys/wait.h>
 #include <unistd.h>
 
+#include "common.h"
+
 #define MEMFD_STR	"memfd:"
+#define MEMFD_HUGE_STR	"memfd-hugetlb:"
 #define SHARED_FT_STR	"(shared file-table)"
 
 #define MFD_DEF_SIZE 8192
@@ -28,41 +31,8 @@
 /*
  * Default is not to test hugetlbfs
  */
-static int hugetlbfs_test;
 static size_t mfd_def_size = MFD_DEF_SIZE;
-
-/*
- * Copied from mlock2-tests.c
- */
-static unsigned long default_huge_page_size(void)
-{
-	unsigned long hps = 0;
-	char *line = NULL;
-	size_t linelen = 0;
-	FILE *f = fopen("/proc/meminfo", "r");
-
-	if (!f)
-		return 0;
-	while (getline(&line, &linelen, f) > 0) {
-		if (sscanf(line, "Hugepagesize:       %lu kB", &hps) == 1) {
-			hps <<= 10;
-			break;
-		}
-	}
-
-	free(line);
-	fclose(f);
-	return hps;
-}
-
-static int sys_memfd_create(const char *name,
-			    unsigned int flags)
-{
-	if (hugetlbfs_test)
-		flags |= MFD_HUGETLB;
-
-	return syscall(__NR_memfd_create, name, flags);
-}
+static const char *memfd_str = MEMFD_STR;
 
 static int mfd_assert_new(const char *name, loff_t sz, unsigned int flags)
 {
@@ -513,6 +483,10 @@ static void mfd_assert_grow_write(int fd)
 	static char *buf;
 	ssize_t l;
 
+	/* hugetlbfs does not support write */
+	if (hugetlbfs_test)
+		return;
+
 	buf = malloc(mfd_def_size * 8);
 	if (!buf) {
 		printf("malloc(%zu) failed: %m\n", mfd_def_size * 8);
@@ -533,6 +507,10 @@ static void mfd_fail_grow_write(int fd)
 	static char *buf;
 	ssize_t l;
 
+	/* hugetlbfs does not support write */
+	if (hugetlbfs_test)
+		return;
+
 	buf = malloc(mfd_def_size * 8);
 	if (!buf) {
 		printf("malloc(%zu) failed: %m\n", mfd_def_size * 8);
@@ -598,7 +576,7 @@ static void test_create(void)
 	char buf[2048];
 	int fd;
 
-	printf("%s CREATE\n", MEMFD_STR);
+	printf("%s CREATE\n", memfd_str);
 
 	/* test NULL name */
 	mfd_fail_new(NULL, 0);
@@ -627,18 +605,13 @@ static void test_create(void)
 	fd = mfd_assert_new("", 0, MFD_CLOEXEC);
 	close(fd);
 
-	if (!hugetlbfs_test) {
-		/* verify MFD_ALLOW_SEALING is allowed */
-		fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING);
-		close(fd);
-
-		/* verify MFD_ALLOW_SEALING | MFD_CLOEXEC is allowed */
-		fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING | MFD_CLOEXEC);
-		close(fd);
-	} else {
-		/* sealing is not supported on hugetlbfs */
-		mfd_fail_new("", MFD_ALLOW_SEALING);
-	}
+	/* verify MFD_ALLOW_SEALING is allowed */
+	fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING);
+	close(fd);
+
+	/* verify MFD_ALLOW_SEALING | MFD_CLOEXEC is allowed */
+	fd = mfd_assert_new("", 0, MFD_ALLOW_SEALING | MFD_CLOEXEC);
+	close(fd);
 }
 
 /*
@@ -649,11 +622,7 @@ static void test_basic(void)
 {
 	int fd;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s BASIC\n", MEMFD_STR);
+	printf("%s BASIC\n", memfd_str);
 
 	fd = mfd_assert_new("kern_memfd_basic",
 			    mfd_def_size,
@@ -698,28 +667,6 @@ static void test_basic(void)
 }
 
 /*
- * hugetlbfs doesn't support seals or write, so just verify grow and shrink
- * on a hugetlbfs file created via memfd_create.
- */
-static void test_hugetlbfs_grow_shrink(void)
-{
-	int fd;
-
-	printf("%s HUGETLBFS-GROW-SHRINK\n", MEMFD_STR);
-
-	fd = mfd_assert_new("kern_memfd_seal_write",
-			    mfd_def_size,
-			    MFD_CLOEXEC);
-
-	mfd_assert_read(fd);
-	mfd_assert_write(fd);
-	mfd_assert_shrink(fd);
-	mfd_assert_grow(fd);
-
-	close(fd);
-}
-
-/*
  * Test SEAL_WRITE
  * Test whether SEAL_WRITE actually prevents modifications.
  */
@@ -727,14 +674,7 @@ static void test_seal_write(void)
 {
 	int fd;
 
-	/*
-	 * hugetlbfs does not contain sealing or write support.  Just test
-	 * basic grow and shrink via test_hugetlbfs_grow_shrink.
-	 */
-	if (hugetlbfs_test)
-		return test_hugetlbfs_grow_shrink();
-
-	printf("%s SEAL-WRITE\n", MEMFD_STR);
+	printf("%s SEAL-WRITE\n", memfd_str);
 
 	fd = mfd_assert_new("kern_memfd_seal_write",
 			    mfd_def_size,
@@ -760,11 +700,7 @@ static void test_seal_shrink(void)
 {
 	int fd;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s SEAL-SHRINK\n", MEMFD_STR);
+	printf("%s SEAL-SHRINK\n", memfd_str);
 
 	fd = mfd_assert_new("kern_memfd_seal_shrink",
 			    mfd_def_size,
@@ -790,11 +726,7 @@ static void test_seal_grow(void)
 {
 	int fd;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s SEAL-GROW\n", MEMFD_STR);
+	printf("%s SEAL-GROW\n", memfd_str);
 
 	fd = mfd_assert_new("kern_memfd_seal_grow",
 			    mfd_def_size,
@@ -820,11 +752,7 @@ static void test_seal_resize(void)
 {
 	int fd;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s SEAL-RESIZE\n", MEMFD_STR);
+	printf("%s SEAL-RESIZE\n", memfd_str);
 
 	fd = mfd_assert_new("kern_memfd_seal_resize",
 			    mfd_def_size,
@@ -843,32 +771,6 @@ static void test_seal_resize(void)
 }
 
 /*
- * hugetlbfs does not support seals.  Basic test to dup the memfd created
- * fd and perform some basic operations on it.
- */
-static void hugetlbfs_dup(char *b_suffix)
-{
-	int fd, fd2;
-
-	printf("%s HUGETLBFS-DUP %s\n", MEMFD_STR, b_suffix);
-
-	fd = mfd_assert_new("kern_memfd_share_dup",
-			    mfd_def_size,
-			    MFD_CLOEXEC);
-
-	fd2 = mfd_assert_dup(fd);
-
-	mfd_assert_read(fd);
-	mfd_assert_write(fd);
-
-	mfd_assert_shrink(fd2);
-	mfd_assert_grow(fd2);
-
-	close(fd2);
-	close(fd);
-}
-
-/*
  * Test sharing via dup()
  * Test that seals are shared between dupped FDs and they're all equal.
  */
@@ -876,16 +778,7 @@ static void test_share_dup(char *banner, char *b_suffix)
 {
 	int fd, fd2;
 
-	/*
-	 * hugetlbfs does not contain sealing support.  Perform some
-	 * basic testing on dup'ed fd instead via hugetlbfs_dup.
-	 */
-	if (hugetlbfs_test) {
-		hugetlbfs_dup(b_suffix);
-		return;
-	}
-
-	printf("%s %s %s\n", MEMFD_STR, banner, b_suffix);
+	printf("%s %s %s\n", memfd_str, banner, b_suffix);
 
 	fd = mfd_assert_new("kern_memfd_share_dup",
 			    mfd_def_size,
@@ -927,11 +820,7 @@ static void test_share_mmap(char *banner, char *b_suffix)
 	int fd;
 	void *p;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s %s %s\n", MEMFD_STR,  banner, b_suffix);
+	printf("%s %s %s\n", memfd_str,  banner, b_suffix);
 
 	fd = mfd_assert_new("kern_memfd_share_mmap",
 			    mfd_def_size,
@@ -956,32 +845,6 @@ static void test_share_mmap(char *banner, char *b_suffix)
 }
 
 /*
- * Basic test to make sure we can open the hugetlbfs fd via /proc and
- * perform some simple operations on it.
- */
-static void hugetlbfs_proc_open(char *b_suffix)
-{
-	int fd, fd2;
-
-	printf("%s HUGETLBFS-PROC-OPEN %s\n", MEMFD_STR, b_suffix);
-
-	fd = mfd_assert_new("kern_memfd_share_open",
-			    mfd_def_size,
-			    MFD_CLOEXEC);
-
-	fd2 = mfd_assert_open(fd, O_RDWR, 0);
-
-	mfd_assert_read(fd);
-	mfd_assert_write(fd);
-
-	mfd_assert_shrink(fd2);
-	mfd_assert_grow(fd2);
-
-	close(fd2);
-	close(fd);
-}
-
-/*
  * Test sealing with open(/proc/self/fd/%d)
  * Via /proc we can get access to a separate file-context for the same memfd.
  * This is *not* like dup(), but like a real separate open(). Make sure the
@@ -991,16 +854,7 @@ static void test_share_open(char *banner, char *b_suffix)
 {
 	int fd, fd2;
 
-	/*
-	 * hugetlbfs does not contain sealing support.  So test basic
-	 * functionality of using /proc fd via hugetlbfs_proc_open
-	 */
-	if (hugetlbfs_test) {
-		hugetlbfs_proc_open(b_suffix);
-		return;
-	}
-
-	printf("%s %s %s\n", MEMFD_STR, banner, b_suffix);
+	printf("%s %s %s\n", memfd_str, banner, b_suffix);
 
 	fd = mfd_assert_new("kern_memfd_share_open",
 			    mfd_def_size,
@@ -1043,11 +897,7 @@ static void test_share_fork(char *banner, char *b_suffix)
 	int fd;
 	pid_t pid;
 
-	/* hugetlbfs does not contain sealing support */
-	if (hugetlbfs_test)
-		return;
-
-	printf("%s %s %s\n", MEMFD_STR, banner, b_suffix);
+	printf("%s %s %s\n", memfd_str, banner, b_suffix);
 
 	fd = mfd_assert_new("kern_memfd_share_fork",
 			    mfd_def_size,
@@ -1083,7 +933,11 @@ int main(int argc, char **argv)
 			}
 
 			hugetlbfs_test = 1;
+			memfd_str = MEMFD_HUGE_STR;
 			mfd_def_size = hpage_size * 2;
+		} else {
+			printf("Unknown option: %s\n", argv[1]);
+			abort();
 		}
 	}
 
diff --git a/tools/testing/selftests/memfd/run_fuse_test.sh b/tools/testing/selftests/memfd/run_fuse_test.sh
index 407df68dfe27..22e572e2d66a 100755
--- a/tools/testing/selftests/memfd/run_fuse_test.sh
+++ b/tools/testing/selftests/memfd/run_fuse_test.sh
@@ -10,6 +10,6 @@ set -e
 
 mkdir mnt
 ./fuse_mnt ./mnt
-./fuse_test ./mnt/memfd
+./fuse_test ./mnt/memfd $@
 fusermount -u ./mnt
 rmdir ./mnt
diff --git a/tools/testing/selftests/memfd/run_tests.sh b/tools/testing/selftests/memfd/run_tests.sh
index daabb350697c..c2d41ed81b24 100755
--- a/tools/testing/selftests/memfd/run_tests.sh
+++ b/tools/testing/selftests/memfd/run_tests.sh
@@ -60,6 +60,7 @@ fi
 # Run the hugetlbfs test
 #
 ./memfd_test hugetlbfs
+./run_fuse_test.sh hugetlbfs
 
 #
 # Give back any huge pages allocated for the test
diff --git a/tools/testing/selftests/vm/Makefile b/tools/testing/selftests/vm/Makefile
index 7f45806bd863..fdefa2295ddc 100644
--- a/tools/testing/selftests/vm/Makefile
+++ b/tools/testing/selftests/vm/Makefile
@@ -8,17 +8,18 @@ endif
 CFLAGS = -Wall -I ../../../../usr/include $(EXTRA_CFLAGS)
 LDLIBS = -lrt
 TEST_GEN_FILES = compaction_test
+TEST_GEN_FILES += gup_benchmark
 TEST_GEN_FILES += hugepage-mmap
 TEST_GEN_FILES += hugepage-shm
 TEST_GEN_FILES += map_hugetlb
+TEST_GEN_FILES += mlock-random-test
 TEST_GEN_FILES += mlock2-tests
 TEST_GEN_FILES += on-fault-limit
 TEST_GEN_FILES += thuge-gen
 TEST_GEN_FILES += transhuge-stress
 TEST_GEN_FILES += userfaultfd
-TEST_GEN_FILES += mlock-random-test
+TEST_GEN_FILES += va_128TBswitch
 TEST_GEN_FILES += virtual_address_range
-TEST_GEN_FILES += gup_benchmark
 
 TEST_PROGS := run_vmtests
 
diff --git a/tools/testing/selftests/vm/run_vmtests b/tools/testing/selftests/vm/run_vmtests
index cc826326de87..d2561895a021 100755
--- a/tools/testing/selftests/vm/run_vmtests
+++ b/tools/testing/selftests/vm/run_vmtests
@@ -177,4 +177,15 @@ else
 	echo "[PASS]"
 fi
 
+echo "-----------------------------"
+echo "running virtual address 128TB switch test"
+echo "-----------------------------"
+./va_128TBswitch
+if [ $? -ne 0 ]; then
+    echo "[FAIL]"
+    exitcode=1
+else
+    echo "[PASS]"
+fi
+
 exit $exitcode
diff --git a/tools/testing/selftests/vm/va_128TBswitch.c b/tools/testing/selftests/vm/va_128TBswitch.c
new file mode 100644
index 000000000000..e7fe734c374f
--- /dev/null
+++ b/tools/testing/selftests/vm/va_128TBswitch.c
@@ -0,0 +1,297 @@
+/*
+ *
+ * Authors: Kirill A. Shutemov <kirill.shutemov@linux.intel.com>
+ * Authors: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License, version 2, as
+ * published by the Free Software Foundation.
+
+ * This program is distributed in the hope that it would be useful, but
+ * WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
+ *
+ */
+
+#include <stdio.h>
+#include <sys/mman.h>
+#include <string.h>
+
+#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
+
+#ifdef __powerpc64__
+#define PAGE_SIZE	(64 << 10)
+/*
+ * This will work with 16M and 2M hugepage size
+ */
+#define HUGETLB_SIZE	(16 << 20)
+#else
+#define PAGE_SIZE	(4 << 10)
+#define HUGETLB_SIZE	(2 << 20)
+#endif
+
+/*
+ * >= 128TB is the hint addr value we used to select
+ * large address space.
+ */
+#define ADDR_SWITCH_HINT (1UL << 47)
+#define LOW_ADDR	((void *) (1UL << 30))
+#define HIGH_ADDR	((void *) (1UL << 48))
+
+struct testcase {
+	void *addr;
+	unsigned long size;
+	unsigned long flags;
+	const char *msg;
+	unsigned int low_addr_required:1;
+	unsigned int keep_mapped:1;
+};
+
+static struct testcase testcases[] = {
+	{
+		/*
+		 * If stack is moved, we could possibly allocate
+		 * this at the requested address.
+		 */
+		.addr = ((void *)(ADDR_SWITCH_HINT - PAGE_SIZE)),
+		.size = PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE, PAGE_SIZE)",
+		.low_addr_required = 1,
+	},
+	{
+		/*
+		 * We should never allocate at the requested address or above it
+		 * The len cross the 128TB boundary. Without MAP_FIXED
+		 * we will always search in the lower address space.
+		 */
+		.addr = ((void *)(ADDR_SWITCH_HINT - PAGE_SIZE)),
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE, (2 * PAGE_SIZE))",
+		.low_addr_required = 1,
+	},
+	{
+		/*
+		 * Exact mapping at 128TB, the area is free we should get that
+		 * even without MAP_FIXED.
+		 */
+		.addr = ((void *)(ADDR_SWITCH_HINT)),
+		.size = PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT, PAGE_SIZE)",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT),
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
+		.msg = "mmap(ADDR_SWITCH_HINT, 2 * PAGE_SIZE, MAP_FIXED)",
+	},
+	{
+		.addr = NULL,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(NULL)",
+		.low_addr_required = 1,
+	},
+	{
+		.addr = LOW_ADDR,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(LOW_ADDR)",
+		.low_addr_required = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(HIGH_ADDR)",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(HIGH_ADDR) again",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
+		.msg = "mmap(HIGH_ADDR, MAP_FIXED)",
+	},
+	{
+		.addr = (void *) -1,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(-1)",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = (void *) -1,
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(-1) again",
+	},
+	{
+		.addr = ((void *)(ADDR_SWITCH_HINT - PAGE_SIZE)),
+		.size = PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE, PAGE_SIZE)",
+		.low_addr_required = 1,
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT - PAGE_SIZE),
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE, 2 * PAGE_SIZE)",
+		.low_addr_required = 1,
+		.keep_mapped = 1,
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT - PAGE_SIZE / 2),
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE/2 , 2 * PAGE_SIZE)",
+		.low_addr_required = 1,
+		.keep_mapped = 1,
+	},
+	{
+		.addr = ((void *)(ADDR_SWITCH_HINT)),
+		.size = PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT, PAGE_SIZE)",
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT),
+		.size = 2 * PAGE_SIZE,
+		.flags = MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
+		.msg = "mmap(ADDR_SWITCH_HINT, 2 * PAGE_SIZE, MAP_FIXED)",
+	},
+};
+
+static struct testcase hugetlb_testcases[] = {
+	{
+		.addr = NULL,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(NULL, MAP_HUGETLB)",
+		.low_addr_required = 1,
+	},
+	{
+		.addr = LOW_ADDR,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(LOW_ADDR, MAP_HUGETLB)",
+		.low_addr_required = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(HIGH_ADDR, MAP_HUGETLB)",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(HIGH_ADDR, MAP_HUGETLB) again",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = HIGH_ADDR,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
+		.msg = "mmap(HIGH_ADDR, MAP_FIXED | MAP_HUGETLB)",
+	},
+	{
+		.addr = (void *) -1,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(-1, MAP_HUGETLB)",
+		.keep_mapped = 1,
+	},
+	{
+		.addr = (void *) -1,
+		.size = HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(-1, MAP_HUGETLB) again",
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT - PAGE_SIZE),
+		.size = 2 * HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
+		.msg = "mmap(ADDR_SWITCH_HINT - PAGE_SIZE, 2*HUGETLB_SIZE, MAP_HUGETLB)",
+		.low_addr_required = 1,
+		.keep_mapped = 1,
+	},
+	{
+		.addr = (void *)(ADDR_SWITCH_HINT),
+		.size = 2 * HUGETLB_SIZE,
+		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
+		.msg = "mmap(ADDR_SWITCH_HINT , 2*HUGETLB_SIZE, MAP_FIXED | MAP_HUGETLB)",
+	},
+};
+
+static int run_test(struct testcase *test, int count)
+{
+	void *p;
+	int i, ret = 0;
+
+	for (i = 0; i < count; i++) {
+		struct testcase *t = test + i;
+
+		p = mmap(t->addr, t->size, PROT_READ | PROT_WRITE, t->flags, -1, 0);
+
+		printf("%s: %p - ", t->msg, p);
+
+		if (p == MAP_FAILED) {
+			printf("FAILED\n");
+			ret = 1;
+			continue;
+		}
+
+		if (t->low_addr_required && p >= (void *)(ADDR_SWITCH_HINT)) {
+			printf("FAILED\n");
+			ret = 1;
+		} else {
+			/*
+			 * Do a dereference of the address returned so that we catch
+			 * bugs in page fault handling
+			 */
+			memset(p, 0, t->size);
+			printf("OK\n");
+		}
+		if (!t->keep_mapped)
+			munmap(p, t->size);
+	}
+
+	return ret;
+}
+
+static int supported_arch(void)
+{
+#if defined(__powerpc64__)
+	return 1;
+#elif defined(__x86_64__)
+	return 1;
+#else
+	return 0;
+#endif
+}
+
+int main(int argc, char **argv)
+{
+	int ret;
+
+	if (!supported_arch())
+		return 0;
+
+	ret = run_test(testcases, ARRAY_SIZE(testcases));
+	if (argc == 2 && !strcmp(argv[1], "--run-hugetlb"))
+		ret = run_test(hugetlb_testcases, ARRAY_SIZE(hugetlb_testcases));
+	return ret;
+}
diff --git a/tools/testing/selftests/x86/5lvl.c b/tools/testing/selftests/x86/5lvl.c
deleted file mode 100644
index 2eafdcd4c2b3..000000000000
--- a/tools/testing/selftests/x86/5lvl.c
+++ /dev/null
@@ -1,177 +0,0 @@
-#include <stdio.h>
-#include <sys/mman.h>
-
-#define ARRAY_SIZE(arr) (sizeof(arr) / sizeof((arr)[0]))
-
-#define PAGE_SIZE	4096
-#define LOW_ADDR	((void *) (1UL << 30))
-#define HIGH_ADDR	((void *) (1UL << 50))
-
-struct testcase {
-	void *addr;
-	unsigned long size;
-	unsigned long flags;
-	const char *msg;
-	unsigned int low_addr_required:1;
-	unsigned int keep_mapped:1;
-};
-
-static struct testcase testcases[] = {
-	{
-		.addr = NULL,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(NULL)",
-		.low_addr_required = 1,
-	},
-	{
-		.addr = LOW_ADDR,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(LOW_ADDR)",
-		.low_addr_required = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(HIGH_ADDR)",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(HIGH_ADDR) again",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
-		.msg = "mmap(HIGH_ADDR, MAP_FIXED)",
-	},
-	{
-		.addr = (void*) -1,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(-1)",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = (void*) -1,
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(-1) again",
-	},
-	{
-		.addr = (void *)((1UL << 47) - PAGE_SIZE),
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap((1UL << 47), 2 * PAGE_SIZE)",
-		.low_addr_required = 1,
-		.keep_mapped = 1,
-	},
-	{
-		.addr = (void *)((1UL << 47) - PAGE_SIZE / 2),
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap((1UL << 47), 2 * PAGE_SIZE / 2)",
-		.low_addr_required = 1,
-		.keep_mapped = 1,
-	},
-	{
-		.addr = (void *)((1UL << 47) - PAGE_SIZE),
-		.size = 2 * PAGE_SIZE,
-		.flags = MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
-		.msg = "mmap((1UL << 47) - PAGE_SIZE, 2 * PAGE_SIZE, MAP_FIXED)",
-	},
-	{
-		.addr = NULL,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(NULL, MAP_HUGETLB)",
-		.low_addr_required = 1,
-	},
-	{
-		.addr = LOW_ADDR,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(LOW_ADDR, MAP_HUGETLB)",
-		.low_addr_required = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(HIGH_ADDR, MAP_HUGETLB)",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(HIGH_ADDR, MAP_HUGETLB) again",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = HIGH_ADDR,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
-		.msg = "mmap(HIGH_ADDR, MAP_FIXED | MAP_HUGETLB)",
-	},
-	{
-		.addr = (void*) -1,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(-1, MAP_HUGETLB)",
-		.keep_mapped = 1,
-	},
-	{
-		.addr = (void*) -1,
-		.size = 2UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap(-1, MAP_HUGETLB) again",
-	},
-	{
-		.addr = (void *)((1UL << 47) - PAGE_SIZE),
-		.size = 4UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS,
-		.msg = "mmap((1UL << 47), 4UL << 20, MAP_HUGETLB)",
-		.low_addr_required = 1,
-		.keep_mapped = 1,
-	},
-	{
-		.addr = (void *)((1UL << 47) - (2UL << 20)),
-		.size = 4UL << 20,
-		.flags = MAP_HUGETLB | MAP_PRIVATE | MAP_ANONYMOUS | MAP_FIXED,
-		.msg = "mmap((1UL << 47) - (2UL << 20), 4UL << 20, MAP_FIXED | MAP_HUGETLB)",
-	},
-};
-
-int main(int argc, char **argv)
-{
-	int i;
-	void *p;
-
-	for (i = 0; i < ARRAY_SIZE(testcases); i++) {
-		struct testcase *t = testcases + i;
-
-		p = mmap(t->addr, t->size, PROT_NONE, t->flags, -1, 0);
-
-		printf("%s: %p - ", t->msg, p);
-
-		if (p == MAP_FAILED) {
-			printf("FAILED\n");
-			continue;
-		}
-
-		if (t->low_addr_required && p >= (void *)(1UL << 47))
-			printf("FAILED\n");
-		else
-			printf("OK\n");
-		if (!t->keep_mapped)
-			munmap(p, t->size);
-	}
-	return 0;
-}
diff --git a/tools/vm/page-types.c b/tools/vm/page-types.c
index e92903fc7113..a8783f48f77f 100644
--- a/tools/vm/page-types.c
+++ b/tools/vm/page-types.c
@@ -169,9 +169,10 @@ static int		opt_raw;	/* for kernel developers */
 static int		opt_list;	/* list pages (in ranges) */
 static int		opt_no_summary;	/* don't show summary */
 static pid_t		opt_pid;	/* process to walk */
-const char *		opt_file;	/* file or directory path */
+const char		*opt_file;	/* file or directory path */
 static uint64_t		opt_cgroup;	/* cgroup inode */
 static int		opt_list_cgroup;/* list page cgroup */
+static const char	*opt_kpageflags;/* kpageflags file to parse */
 
 #define MAX_ADDR_RANGES	1024
 static int		nr_addr_ranges;
@@ -258,7 +259,7 @@ static int checked_open(const char *pathname, int flags)
  * pagemap/kpageflags routines
  */
 
-static unsigned long do_u64_read(int fd, char *name,
+static unsigned long do_u64_read(int fd, const char *name,
 				 uint64_t *buf,
 				 unsigned long index,
 				 unsigned long count)
@@ -283,7 +284,7 @@ static unsigned long kpageflags_read(uint64_t *buf,
 				     unsigned long index,
 				     unsigned long pages)
 {
-	return do_u64_read(kpageflags_fd, PROC_KPAGEFLAGS, buf, index, pages);
+	return do_u64_read(kpageflags_fd, opt_kpageflags, buf, index, pages);
 }
 
 static unsigned long kpagecgroup_read(uint64_t *buf,
@@ -293,7 +294,7 @@ static unsigned long kpagecgroup_read(uint64_t *buf,
 	if (kpagecgroup_fd < 0)
 		return pages;
 
-	return do_u64_read(kpagecgroup_fd, PROC_KPAGEFLAGS, buf, index, pages);
+	return do_u64_read(kpagecgroup_fd, opt_kpageflags, buf, index, pages);
 }
 
 static unsigned long pagemap_read(uint64_t *buf,
@@ -743,7 +744,7 @@ static void walk_addr_ranges(void)
 {
 	int i;
 
-	kpageflags_fd = checked_open(PROC_KPAGEFLAGS, O_RDONLY);
+	kpageflags_fd = checked_open(opt_kpageflags, O_RDONLY);
 
 	if (!nr_addr_ranges)
 		add_addr_range(0, ULONG_MAX);
@@ -790,6 +791,7 @@ static void usage(void)
 "            -N|--no-summary            Don't show summary info\n"
 "            -X|--hwpoison              hwpoison pages\n"
 "            -x|--unpoison              unpoison pages\n"
+"            -F|--kpageflags filename   kpageflags file to parse\n"
 "            -h|--help                  Show this usage message\n"
 "flags:\n"
 "            0x10                       bitfield format, e.g.\n"
@@ -1013,7 +1015,7 @@ static void walk_page_cache(void)
 {
 	struct stat st;
 
-	kpageflags_fd = checked_open(PROC_KPAGEFLAGS, O_RDONLY);
+	kpageflags_fd = checked_open(opt_kpageflags, O_RDONLY);
 	pagemap_fd = checked_open("/proc/self/pagemap", O_RDONLY);
 	sigaction(SIGBUS, &sigbus_action, NULL);
 
@@ -1164,6 +1166,11 @@ static void parse_bits_mask(const char *optarg)
 	add_bits_filter(mask, bits);
 }
 
+static void parse_kpageflags(const char *name)
+{
+	opt_kpageflags = name;
+}
+
 static void describe_flags(const char *optarg)
 {
 	uint64_t flags = parse_flag_names(optarg, 0);
@@ -1188,6 +1195,7 @@ static const struct option opts[] = {
 	{ "no-summary", 0, NULL, 'N' },
 	{ "hwpoison"  , 0, NULL, 'X' },
 	{ "unpoison"  , 0, NULL, 'x' },
+	{ "kpageflags", 0, NULL, 'F' },
 	{ "help"      , 0, NULL, 'h' },
 	{ NULL        , 0, NULL, 0 }
 };
@@ -1199,7 +1207,7 @@ int main(int argc, char *argv[])
 	page_size = getpagesize();
 
 	while ((c = getopt_long(argc, argv,
-				"rp:f:a:b:d:c:ClLNXxh", opts, NULL)) != -1) {
+				"rp:f:a:b:d:c:ClLNXxF:h", opts, NULL)) != -1) {
 		switch (c) {
 		case 'r':
 			opt_raw = 1;
@@ -1242,6 +1250,9 @@ int main(int argc, char *argv[])
 			opt_unpoison = 1;
 			prepare_hwpoison_fd();
 			break;
+		case 'F':
+			parse_kpageflags(optarg);
+			break;
 		case 'h':
 			usage();
 			exit(0);
@@ -1251,6 +1262,9 @@ int main(int argc, char *argv[])
 		}
 	}
 
+	if (!opt_kpageflags)
+		opt_kpageflags = PROC_KPAGEFLAGS;
+
 	if (opt_cgroup || opt_list_cgroup)
 		kpagecgroup_fd = checked_open(PROC_KPAGECGROUP, O_RDONLY);