14 files changed, 393 insertions, 75 deletions

diff --git a/tools/hv/hv_kvp_daemon.c b/tools/hv/hv_kvp_daemon.c
index 5959affd8820..d25a46925e61 100644
--- a/tools/hv/hv_kvp_daemon.c
+++ b/tools/hv/hv_kvp_daemon.c
@@ -43,6 +43,7 @@
 #include <sys/stat.h>
 #include <fcntl.h>
 #include <dirent.h>
+#include <net/if.h>
 
 /*
  * KVP protocol: The user mode component first registers with the
@@ -88,6 +89,7 @@ static char *os_major = "";
 static char *os_minor = "";
 static char *processor_arch;
 static char *os_build;
+static char *os_version;
 static char *lic_version = "Unknown version";
 static struct utsname uts_buf;
 
@@ -297,7 +299,7 @@ static int kvp_file_init(void)
 	return 0;
 }
 
-static int kvp_key_delete(int pool, __u8 *key, int key_size)
+static int kvp_key_delete(int pool, const char *key, int key_size)
 {
 	int i;
 	int j, k;
@@ -340,7 +342,7 @@ static int kvp_key_delete(int pool, __u8 *key, int key_size)
 	return 1;
 }
 
-static int kvp_key_add_or_modify(int pool, __u8 *key, int key_size, __u8 *value,
+static int kvp_key_add_or_modify(int pool, const char *key, int key_size, const char *value,
 			int value_size)
 {
 	int i;
@@ -394,7 +396,7 @@ static int kvp_key_add_or_modify(int pool, __u8 *key, int key_size, __u8 *value,
 	return 0;
 }
 
-static int kvp_get_value(int pool, __u8 *key, int key_size, __u8 *value,
+static int kvp_get_value(int pool, const char *key, int key_size, char *value,
 			int value_size)
 {
 	int i;
@@ -426,8 +428,8 @@ static int kvp_get_value(int pool, __u8 *key, int key_size, __u8 *value,
 	return 1;
 }
 
-static int kvp_pool_enumerate(int pool, int index, __u8 *key, int key_size,
-				__u8 *value, int value_size)
+static int kvp_pool_enumerate(int pool, int index, char *key, int key_size,
+				char *value, int value_size)
 {
 	struct kvp_record *record;
 
@@ -453,7 +455,9 @@ void kvp_get_os_info(void)
 	char	*p, buf[512];
 
 	uname(&uts_buf);
-	os_build = uts_buf.release;
+	os_version = uts_buf.release;
+	os_build = strdup(uts_buf.release);
+
 	os_name = uts_buf.sysname;
 	processor_arch = uts_buf.machine;
 
@@ -462,7 +466,7 @@ void kvp_get_os_info(void)
 	 * string to be of the form: x.y.z
 	 * Strip additional information we may have.
 	 */
-	p = strchr(os_build, '-');
+	p = strchr(os_version, '-');
 	if (p)
 		*p = '\0';
 
@@ -879,7 +883,7 @@ static int kvp_process_ip_address(void *addrp,
 		addr_length = INET6_ADDRSTRLEN;
 	}
 
-	if ((length - *offset) < addr_length + 1)
+	if ((length - *offset) < addr_length + 2)
 		return HV_E_FAIL;
 	if (str == NULL) {
 		strcpy(buffer, "inet_ntop failed\n");
@@ -887,11 +891,13 @@ static int kvp_process_ip_address(void *addrp,
 	}
 	if (*offset == 0)
 		strcpy(buffer, tmp);
-	else
+	else {
+		strcat(buffer, ";");
 		strcat(buffer, tmp);
-	strcat(buffer, ";");
+	}
 
 	*offset += strlen(str) + 1;
+
 	return 0;
 }
 
@@ -953,7 +959,9 @@ kvp_get_ip_info(int family, char *if_name, int op,
 		 * supported address families; if not we gather info on
 		 * the specified address family.
 		 */
-		if ((family != 0) && (curp->ifa_addr->sa_family != family)) {
+		if ((((family != 0) &&
+			 (curp->ifa_addr->sa_family != family))) ||
+			 (curp->ifa_flags & IFF_LOOPBACK)) {
 			curp = curp->ifa_next;
 			continue;
 		}
@@ -1478,13 +1486,19 @@ int main(void)
 		len = recvfrom(fd, kvp_recv_buffer, sizeof(kvp_recv_buffer), 0,
 				addr_p, &addr_l);
 
-		if (len < 0 || addr.nl_pid) {
+		if (len < 0) {
 			syslog(LOG_ERR, "recvfrom failed; pid:%u error:%d %s",
 					addr.nl_pid, errno, strerror(errno));
 			close(fd);
 			return -1;
 		}
 
+		if (addr.nl_pid) {
+			syslog(LOG_WARNING, "Received packet from untrusted pid:%u",
+					addr.nl_pid);
+			continue;
+		}
+
 		incoming_msg = (struct nlmsghdr *)kvp_recv_buffer;
 		incoming_cn_msg = (struct cn_msg *)NLMSG_DATA(incoming_msg);
 		hv_msg = (struct hv_kvp_msg *)incoming_cn_msg->data;
@@ -1649,7 +1663,7 @@ int main(void)
 			strcpy(key_name, "OSMinorVersion");
 			break;
 		case OSVersion:
-			strcpy(key_value, os_build);
+			strcpy(key_value, os_version);
 			strcpy(key_name, "OSVersion");
 			break;
 		case ProcessorArchitecture:
diff --git a/tools/power/cpupower/.gitignore b/tools/power/cpupower/.gitignore
index 8a83dd2ffc11..d42073f12609 100644
--- a/tools/power/cpupower/.gitignore
+++ b/tools/power/cpupower/.gitignore
@@ -20,3 +20,10 @@ utils/cpufreq-set.o
 utils/cpufreq-aperf.o
 cpupower
 bench/cpufreq-bench
+debug/kernel/Module.symvers
+debug/i386/centrino-decode
+debug/i386/dump_psb
+debug/i386/intel_gsic
+debug/i386/powernow-k8-decode
+debug/x86_64/centrino-decode
+debug/x86_64/powernow-k8-decode
diff --git a/tools/power/cpupower/Makefile b/tools/power/cpupower/Makefile
index cf397bd26d0c..d875a74a3bdf 100644
--- a/tools/power/cpupower/Makefile
+++ b/tools/power/cpupower/Makefile
@@ -253,7 +253,8 @@ clean:
 	 | xargs rm -f
 	-rm -f $(OUTPUT)cpupower
 	-rm -f $(OUTPUT)libcpupower.so*
-	-rm -rf $(OUTPUT)po/*.{gmo,pot}
+	-rm -rf $(OUTPUT)po/*.gmo
+	-rm -rf $(OUTPUT)po/*.pot
 	$(MAKE) -C bench O=$(OUTPUT) clean
 
 
diff --git a/tools/power/cpupower/debug/i386/Makefile b/tools/power/cpupower/debug/i386/Makefile
index 3ba158f0e287..c05cc0ac80c7 100644
--- a/tools/power/cpupower/debug/i386/Makefile
+++ b/tools/power/cpupower/debug/i386/Makefile
@@ -26,7 +26,10 @@ $(OUTPUT)powernow-k8-decode: powernow-k8-decode.c
 all: $(OUTPUT)centrino-decode $(OUTPUT)dump_psb $(OUTPUT)intel_gsic $(OUTPUT)powernow-k8-decode
 
 clean:
-	rm -rf $(OUTPUT){centrino-decode,dump_psb,intel_gsic,powernow-k8-decode}
+	rm -rf $(OUTPUT)centrino-decode
+	rm -rf $(OUTPUT)dump_psb
+	rm -rf $(OUTPUT)intel_gsic
+	rm -rf $(OUTPUT)powernow-k8-decode
 
 install:
 	$(INSTALL) -d $(DESTDIR)${bindir}
diff --git a/tools/power/cpupower/man/cpupower-monitor.1 b/tools/power/cpupower/man/cpupower-monitor.1
index 1141c2073719..e01c35d13b6e 100644
--- a/tools/power/cpupower/man/cpupower-monitor.1
+++ b/tools/power/cpupower/man/cpupower-monitor.1
@@ -7,11 +7,11 @@ cpupower\-monitor \- Report processor frequency and idle statistics
 .RB "\-l"
 
 .B cpupower monitor
-.RB [ "\-m <mon1>," [ "<mon2>,..." ] ]
+.RB [ -c ] [ "\-m <mon1>," [ "<mon2>,..." ] ]
 .RB [ "\-i seconds" ]
 .br
 .B cpupower monitor
-.RB [ "\-m <mon1>," [ "<mon2>,..." ] ]
+.RB [ -c ][ "\-m <mon1>," [ "<mon2>,..." ] ]
 .RB command
 .br
 .SH DESCRIPTION
@@ -64,6 +64,17 @@ Only display specific monitors. Use the monitor string(s) provided by \-l option
 Measure intervall.
 .RE
 .PP
+\-c
+.RS 4
+Schedule the process on every core before starting and ending measuring.
+This could be needed for the Idle_Stats monitor when no other MSR based
+monitor (has to be run on the core that is measured) is run in parallel.
+This is to wake up the processors from deeper sleep states and let the
+kernel re
+-account its cpuidle (C-state) information before reading the
+cpuidle timings from sysfs.
+.RE
+.PP
 command
 .RS 4
 Measure idle and frequency characteristics of an arbitrary command/workload.
diff --git a/tools/power/cpupower/utils/helpers/cpuid.c b/tools/power/cpupower/utils/helpers/cpuid.c
index 906895d21cce..93b0aa74ca03 100644
--- a/tools/power/cpupower/utils/helpers/cpuid.c
+++ b/tools/power/cpupower/utils/helpers/cpuid.c
@@ -158,6 +158,8 @@ out:
 				cpu_info->caps |= CPUPOWER_CAP_HAS_TURBO_RATIO;
 			case 0x2A:	/* SNB */
 			case 0x2D:	/* SNB Xeon */
+			case 0x3A:	/* IVB */
+			case 0x3E:	/* IVB Xeon */
 				cpu_info->caps |= CPUPOWER_CAP_HAS_TURBO_RATIO;
 				cpu_info->caps |= CPUPOWER_CAP_IS_SNB;
 				break;
diff --git a/tools/power/cpupower/utils/helpers/helpers.h b/tools/power/cpupower/utils/helpers/helpers.h
index 2eb584cf2f55..aa9e95486a2d 100644
--- a/tools/power/cpupower/utils/helpers/helpers.h
+++ b/tools/power/cpupower/utils/helpers/helpers.h
@@ -92,6 +92,14 @@ extern int get_cpu_info(unsigned int cpu, struct cpupower_cpu_info *cpu_info);
 extern struct cpupower_cpu_info cpupower_cpu_info;
 /* cpuid and cpuinfo helpers  **************************/
 
+struct cpuid_core_info {
+	int pkg;
+	int core;
+	int cpu;
+
+	/* flags */
+	unsigned int is_online:1;
+};
 
 /* CPU topology/hierarchy parsing ******************/
 struct cpupower_topology {
@@ -101,18 +109,12 @@ struct cpupower_topology {
 	unsigned int threads; /* per core */
 
 	/* Array gets mallocated with cores entries, holding per core info */
-	struct {
-		int pkg;
-		int core;
-		int cpu;
-
-		/* flags */
-		unsigned int is_online:1;
-	} *core_info;
+	struct cpuid_core_info *core_info;
 };
 
 extern int get_cpu_topology(struct cpupower_topology *cpu_top);
 extern void cpu_topology_release(struct cpupower_topology cpu_top);
+
 /* CPU topology/hierarchy parsing ******************/
 
 /* X86 ONLY ****************************************/
diff --git a/tools/power/cpupower/utils/helpers/sysfs.c b/tools/power/cpupower/utils/helpers/sysfs.c
index 96e28c124b5c..38ab91629463 100644
--- a/tools/power/cpupower/utils/helpers/sysfs.c
+++ b/tools/power/cpupower/utils/helpers/sysfs.c
@@ -37,25 +37,6 @@ unsigned int sysfs_read_file(const char *path, char *buf, size_t buflen)
 	return (unsigned int) numread;
 }
 
-static unsigned int sysfs_write_file(const char *path,
-				     const char *value, size_t len)
-{
-	int fd;
-	ssize_t numwrite;
-
-	fd = open(path, O_WRONLY);
-	if (fd == -1)
-		return 0;
-
-	numwrite = write(fd, value, len);
-	if (numwrite < 1) {
-		close(fd);
-		return 0;
-	}
-	close(fd);
-	return (unsigned int) numwrite;
-}
-
 /*
  * Detect whether a CPU is online
  *
diff --git a/tools/power/cpupower/utils/helpers/topology.c b/tools/power/cpupower/utils/helpers/topology.c
index 4eae2c47ba48..c13120af519b 100644
--- a/tools/power/cpupower/utils/helpers/topology.c
+++ b/tools/power/cpupower/utils/helpers/topology.c
@@ -20,9 +20,8 @@
 #include <helpers/sysfs.h>
 
 /* returns -1 on failure, 0 on success */
-int sysfs_topology_read_file(unsigned int cpu, const char *fname)
+static int sysfs_topology_read_file(unsigned int cpu, const char *fname, int *result)
 {
-	unsigned long value;
 	char linebuf[MAX_LINE_LEN];
 	char *endp;
 	char path[SYSFS_PATH_MAX];
@@ -31,20 +30,12 @@ int sysfs_topology_read_file(unsigned int cpu, const char *fname)
 			 cpu, fname);
 	if (sysfs_read_file(path, linebuf, MAX_LINE_LEN) == 0)
 		return -1;
-	value = strtoul(linebuf, &endp, 0);
+	*result = strtol(linebuf, &endp, 0);
 	if (endp == linebuf || errno == ERANGE)
 		return -1;
-	return value;
+	return 0;
 }
 
-struct cpuid_core_info {
-	unsigned int pkg;
-	unsigned int thread;
-	unsigned int cpu;
-	/* flags */
-	unsigned int is_online:1;
-};
-
 static int __compare(const void *t1, const void *t2)
 {
 	struct cpuid_core_info *top1 = (struct cpuid_core_info *)t1;
@@ -53,9 +44,9 @@ static int __compare(const void *t1, const void *t2)
 		return -1;
 	else if (top1->pkg > top2->pkg)
 		return 1;
-	else if (top1->thread < top2->thread)
+	else if (top1->core < top2->core)
 		return -1;
-	else if (top1->thread > top2->thread)
+	else if (top1->core > top2->core)
 		return 1;
 	else if (top1->cpu < top2->cpu)
 		return -1;
@@ -73,28 +64,42 @@ static int __compare(const void *t1, const void *t2)
  */
 int get_cpu_topology(struct cpupower_topology *cpu_top)
 {
-	int cpu, cpus = sysconf(_SC_NPROCESSORS_CONF);
+	int cpu, last_pkg, cpus = sysconf(_SC_NPROCESSORS_CONF);
 
-	cpu_top->core_info = malloc(sizeof(struct cpupower_topology) * cpus);
+	cpu_top->core_info = malloc(sizeof(struct cpuid_core_info) * cpus);
 	if (cpu_top->core_info == NULL)
 		return -ENOMEM;
 	cpu_top->pkgs = cpu_top->cores = 0;
 	for (cpu = 0; cpu < cpus; cpu++) {
 		cpu_top->core_info[cpu].cpu = cpu;
 		cpu_top->core_info[cpu].is_online = sysfs_is_cpu_online(cpu);
-		cpu_top->core_info[cpu].pkg =
-			sysfs_topology_read_file(cpu, "physical_package_id");
-		if ((int)cpu_top->core_info[cpu].pkg != -1 &&
-		    cpu_top->core_info[cpu].pkg > cpu_top->pkgs)
-			cpu_top->pkgs = cpu_top->core_info[cpu].pkg;
-		cpu_top->core_info[cpu].core =
-			sysfs_topology_read_file(cpu, "core_id");
+		if(sysfs_topology_read_file(
+			cpu,
+			"physical_package_id",
+			&(cpu_top->core_info[cpu].pkg)) < 0)
+			return -1;
+		if(sysfs_topology_read_file(
+			cpu,
+			"core_id",
+			&(cpu_top->core_info[cpu].core)) < 0)
+			return -1;
 	}
-	cpu_top->pkgs++;
 
 	qsort(cpu_top->core_info, cpus, sizeof(struct cpuid_core_info),
 	      __compare);
 
+	/* Count the number of distinct pkgs values. This works
+	   because the primary sort of the core_info struct was just
+	   done by pkg value. */
+	last_pkg = cpu_top->core_info[0].pkg;
+	for(cpu = 1; cpu < cpus; cpu++) {
+		if(cpu_top->core_info[cpu].pkg != last_pkg) {
+			last_pkg = cpu_top->core_info[cpu].pkg;
+			cpu_top->pkgs++;
+		}
+	}
+	cpu_top->pkgs++;
+
 	/* Intel's cores count is not consecutively numbered, there may
 	 * be a core_id of 3, but none of 2. Assume there always is 0
 	 * Get amount of cores by counting duplicates in a package
diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c
index 0d6571e418db..c4bae9203a69 100644
--- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c
+++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.c
@@ -39,6 +39,7 @@ static int mode;
 static int interval = 1;
 static char *show_monitors_param;
 static struct cpupower_topology cpu_top;
+static unsigned int wake_cpus;
 
 /* ToDo: Document this in the manpage */
 static char range_abbr[RANGE_MAX] = { 'T', 'C', 'P', 'M', };
@@ -84,7 +85,7 @@ int fill_string_with_spaces(char *s, int n)
 void print_header(int topology_depth)
 {
 	int unsigned mon;
-	int state, need_len, pr_mon_len;
+	int state, need_len;
 	cstate_t s;
 	char buf[128] = "";
 	int percent_width = 4;
@@ -93,7 +94,6 @@ void print_header(int topology_depth)
 	printf("%s|", buf);
 
 	for (mon = 0; mon < avail_monitors; mon++) {
-		pr_mon_len = 0;
 		need_len = monitors[mon]->hw_states_num * (percent_width + 3)
 			- 1;
 		if (mon != 0) {
@@ -315,16 +315,28 @@ int fork_it(char **argv)
 int do_interval_measure(int i)
 {
 	unsigned int num;
+	int cpu;
+
+	if (wake_cpus)
+		for (cpu = 0; cpu < cpu_count; cpu++)
+			bind_cpu(cpu);
 
 	for (num = 0; num < avail_monitors; num++) {
 		dprint("HW C-state residency monitor: %s - States: %d\n",
 		       monitors[num]->name, monitors[num]->hw_states_num);
 		monitors[num]->start();
 	}
+
 	sleep(i);
+
+	if (wake_cpus)
+		for (cpu = 0; cpu < cpu_count; cpu++)
+			bind_cpu(cpu);
+
 	for (num = 0; num < avail_monitors; num++)
 		monitors[num]->stop();
 
+
 	return 0;
 }
 
@@ -333,7 +345,7 @@ static void cmdline(int argc, char *argv[])
 	int opt;
 	progname = basename(argv[0]);
 
-	while ((opt = getopt(argc, argv, "+li:m:")) != -1) {
+	while ((opt = getopt(argc, argv, "+lci:m:")) != -1) {
 		switch (opt) {
 		case 'l':
 			if (mode)
@@ -352,6 +364,9 @@ static void cmdline(int argc, char *argv[])
 			mode = show;
 			show_monitors_param = optarg;
 			break;
+		case 'c':
+			wake_cpus = 1;
+			break;
 		default:
 			print_wrong_arg_exit();
 		}
diff --git a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h
index 9312ee1f2dbc..9e43f3371fbc 100644
--- a/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h
+++ b/tools/power/cpupower/utils/idle_monitor/cpupower-monitor.h
@@ -65,4 +65,21 @@ extern long long timespec_diff_us(struct timespec start, struct timespec end);
 				"could be inaccurate\n"), mes, ov);		\
 }
 
+
+/* Taken over from x86info project sources  -> return 0 on success */
+#include <sched.h>
+#include <sys/types.h>
+#include <unistd.h>
+static inline int bind_cpu(int cpu)
+{
+	cpu_set_t set;
+
+	if (sched_getaffinity(getpid(), sizeof(set), &set) == 0) {
+		CPU_ZERO(&set);
+		CPU_SET(cpu, &set);
+		return sched_setaffinity(getpid(), sizeof(set), &set);
+	}
+	return 1;
+}
+
 #endif /* __CPUIDLE_INFO_HW__ */
diff --git a/tools/power/cpupower/utils/idle_monitor/snb_idle.c b/tools/power/cpupower/utils/idle_monitor/snb_idle.c
index a1bc07cd53e1..a99b43b97d6d 100644
--- a/tools/power/cpupower/utils/idle_monitor/snb_idle.c
+++ b/tools/power/cpupower/utils/idle_monitor/snb_idle.c
@@ -150,9 +150,15 @@ static struct cpuidle_monitor *snb_register(void)
 	    || cpupower_cpu_info.family != 6)
 		return NULL;
 
-	if (cpupower_cpu_info.model != 0x2A
-	    && cpupower_cpu_info.model != 0x2D)
+	switch (cpupower_cpu_info.model) {
+	case 0x2A: /* SNB */
+	case 0x2D: /* SNB Xeon */
+	case 0x3A: /* IVB */
+	case 0x3E: /* IVB Xeon */
+		break;
+	default:
 		return NULL;
+	}
 
 	is_valid = calloc(cpu_count, sizeof(int));
 	for (num = 0; num < SNB_CSTATE_COUNT; num++) {
diff --git a/tools/testing/selftests/vm/Makefile b/tools/testing/selftests/vm/Makefile
index b336b24aa6c0..7300d0702efe 100644
--- a/tools/testing/selftests/vm/Makefile
+++ b/tools/testing/selftests/vm/Makefile
@@ -1,9 +1,9 @@
 # Makefile for vm selftests
 
 CC = $(CROSS_COMPILE)gcc
-CFLAGS = -Wall -Wextra
+CFLAGS = -Wall
 
-all: hugepage-mmap hugepage-shm  map_hugetlb
+all: hugepage-mmap hugepage-shm  map_hugetlb thuge-gen
 %: %.c
 	$(CC) $(CFLAGS) -o $@ $^
 
diff --git a/tools/testing/selftests/vm/thuge-gen.c b/tools/testing/selftests/vm/thuge-gen.c
new file mode 100644
index 000000000000..c87957295f74
--- /dev/null
+++ b/tools/testing/selftests/vm/thuge-gen.c
@@ -0,0 +1,254 @@
+/* Test selecting other page sizes for mmap/shmget.
+
+   Before running this huge pages for each huge page size must have been
+   reserved.
+   For large pages beyond MAX_ORDER (like 1GB on x86) boot options must be used.
+   Also shmmax must be increased.
+   And you need to run as root to work around some weird permissions in shm.
+   And nothing using huge pages should run in parallel.
+   When the program aborts you may need to clean up the shm segments with
+   ipcrm -m by hand, like this
+   sudo ipcs | awk '$1 == "0x00000000" {print $2}' | xargs -n1 sudo ipcrm -m
+   (warning this will remove all if someone else uses them) */
+
+#define _GNU_SOURCE 1
+#include <sys/mman.h>
+#include <stdlib.h>
+#include <stdio.h>
+#include <sys/ipc.h>
+#include <sys/shm.h>
+#include <sys/stat.h>
+#include <glob.h>
+#include <assert.h>
+#include <unistd.h>
+#include <stdarg.h>
+#include <string.h>
+
+#define err(x) perror(x), exit(1)
+
+#define MAP_HUGE_2MB    (21 << MAP_HUGE_SHIFT)
+#define MAP_HUGE_1GB    (30 << MAP_HUGE_SHIFT)
+#define MAP_HUGE_SHIFT  26
+#define MAP_HUGE_MASK   0x3f
+#define MAP_HUGETLB	0x40000
+
+#define SHM_HUGETLB     04000   /* segment will use huge TLB pages */
+#define SHM_HUGE_SHIFT  26
+#define SHM_HUGE_MASK   0x3f
+#define SHM_HUGE_2MB    (21 << SHM_HUGE_SHIFT)
+#define SHM_HUGE_1GB    (30 << SHM_HUGE_SHIFT)
+
+#define NUM_PAGESIZES   5
+
+#define NUM_PAGES 4
+
+#define Dprintf(fmt...) // printf(fmt)
+
+unsigned long page_sizes[NUM_PAGESIZES];
+int num_page_sizes;
+
+int ilog2(unsigned long v)
+{
+	int l = 0;
+	while ((1UL << l) < v)
+		l++;
+	return l;
+}
+
+void find_pagesizes(void)
+{
+	glob_t g;
+	int i;
+	glob("/sys/kernel/mm/hugepages/hugepages-*kB", 0, NULL, &g);
+	assert(g.gl_pathc <= NUM_PAGESIZES);
+	for (i = 0; i < g.gl_pathc; i++) {
+		sscanf(g.gl_pathv[i], "/sys/kernel/mm/hugepages/hugepages-%lukB",
+				&page_sizes[i]);
+		page_sizes[i] <<= 10;
+		printf("Found %luMB\n", page_sizes[i] >> 20);
+	}
+	num_page_sizes = g.gl_pathc;
+	globfree(&g);
+}
+
+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);
+	return hps;
+}
+
+void show(unsigned long ps)
+{
+	char buf[100];
+	if (ps == getpagesize())
+		return;
+	printf("%luMB: ", ps >> 20);
+	fflush(stdout);
+	snprintf(buf, sizeof buf,
+		"cat /sys/kernel/mm/hugepages/hugepages-%lukB/free_hugepages",
+		ps >> 10);
+	system(buf);
+}
+
+unsigned long read_sysfs(int warn, char *fmt, ...)
+{
+	char *line = NULL;
+	size_t linelen = 0;
+	char buf[100];
+	FILE *f;
+	va_list ap;
+	unsigned long val = 0;
+
+	va_start(ap, fmt);
+	vsnprintf(buf, sizeof buf, fmt, ap);
+	va_end(ap);
+
+	f = fopen(buf, "r");
+	if (!f) {
+		if (warn)
+			printf("missing %s\n", buf);
+		return 0;
+	}
+	if (getline(&line, &linelen, f) > 0) {
+		sscanf(line, "%lu", &val);
+	}
+	fclose(f);
+	free(line);
+	return val;
+}
+
+unsigned long read_free(unsigned long ps)
+{
+	return read_sysfs(ps != getpagesize(),
+			"/sys/kernel/mm/hugepages/hugepages-%lukB/free_hugepages",
+			ps >> 10);
+}
+
+void test_mmap(unsigned long size, unsigned flags)
+{
+	char *map;
+	unsigned long before, after;
+	int err;
+
+	before = read_free(size);
+	map = mmap(NULL, size*NUM_PAGES, PROT_READ|PROT_WRITE,
+			MAP_PRIVATE|MAP_ANONYMOUS|MAP_HUGETLB|flags, 0, 0);
+
+	if (map == (char *)-1) err("mmap");
+	memset(map, 0xff, size*NUM_PAGES);
+	after = read_free(size);
+	Dprintf("before %lu after %lu diff %ld size %lu\n",
+		before, after, before - after, size);
+	assert(size == getpagesize() || (before - after) == NUM_PAGES);
+	show(size);
+	err = munmap(map, size);
+	assert(!err);
+}
+
+void test_shmget(unsigned long size, unsigned flags)
+{
+	int id;
+	unsigned long before, after;
+	int err;
+
+	before = read_free(size);
+	id = shmget(IPC_PRIVATE, size * NUM_PAGES, IPC_CREAT|0600|flags);
+	if (id < 0) err("shmget");
+
+	struct shm_info i;
+	if (shmctl(id, SHM_INFO, (void *)&i) < 0) err("shmctl");
+	Dprintf("alloc %lu res %lu\n", i.shm_tot, i.shm_rss);
+
+
+	Dprintf("id %d\n", id);
+	char *map = shmat(id, NULL, 0600);
+	if (map == (char*)-1) err("shmat");
+
+	shmctl(id, IPC_RMID, NULL);
+
+	memset(map, 0xff, size*NUM_PAGES);
+	after = read_free(size);
+
+	Dprintf("before %lu after %lu diff %ld size %lu\n",
+		before, after, before - after, size);
+	assert(size == getpagesize() || (before - after) == NUM_PAGES);
+	show(size);
+	err = shmdt(map);
+	assert(!err);
+}
+
+void sanity_checks(void)
+{
+	int i;
+	unsigned long largest = getpagesize();
+
+	for (i = 0; i < num_page_sizes; i++) {
+		if (page_sizes[i] > largest)
+			largest = page_sizes[i];
+
+		if (read_free(page_sizes[i]) < NUM_PAGES) {
+			printf("Not enough huge pages for page size %lu MB, need %u\n",
+				page_sizes[i] >> 20,
+				NUM_PAGES);
+			exit(0);
+		}
+	}
+
+	if (read_sysfs(0, "/proc/sys/kernel/shmmax") < NUM_PAGES * largest) {
+		printf("Please do echo %lu > /proc/sys/kernel/shmmax", largest * NUM_PAGES);
+		exit(0);
+	}
+
+#if defined(__x86_64__)
+	if (largest != 1U<<30) {
+		printf("No GB pages available on x86-64\n"
+		       "Please boot with hugepagesz=1G hugepages=%d\n", NUM_PAGES);
+		exit(0);
+	}
+#endif
+}
+
+int main(void)
+{
+	int i;
+	unsigned default_hps = default_huge_page_size();
+
+	find_pagesizes();
+
+	sanity_checks();
+
+	for (i = 0; i < num_page_sizes; i++) {
+		unsigned long ps = page_sizes[i];
+		int arg = ilog2(ps) << MAP_HUGE_SHIFT;
+		printf("Testing %luMB mmap with shift %x\n", ps >> 20, arg);
+		test_mmap(ps, MAP_HUGETLB | arg);
+	}
+	printf("Testing default huge mmap\n");
+	test_mmap(default_hps, SHM_HUGETLB);
+
+	puts("Testing non-huge shmget");
+	test_shmget(getpagesize(), 0);
+
+	for (i = 0; i < num_page_sizes; i++) {
+		unsigned long ps = page_sizes[i];
+		int arg = ilog2(ps) << SHM_HUGE_SHIFT;
+		printf("Testing %luMB shmget with shift %x\n", ps >> 20, arg);
+		test_shmget(ps, SHM_HUGETLB | arg);
+	}
+	puts("default huge shmget");
+	test_shmget(default_hps, SHM_HUGETLB);
+
+	return 0;
+}