Merge tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-tracegrafted

Pull tracing fixes from Steven Rostedt:

 - Don't destroy user event fields when removal fails

   User event fields are destroyed before the event is removed from
   visibility. But that can fail leaving the still visible event with no
   fields. Move the destroying of the fields to after the event is
   successfully removed from visibility.

 - Initialize function graph state is fork before calling
   copy_exec_state()

   For non-CLONE_VM forks, copy_exec_state() allocates a new
   task_exec_state. If that allocation fails, ftrace_graph_exit_task()
   will free the tasks ret_stack pointer. Since that pointer is still
   using the parent's ret_stack, it mistakenly frees the parent's
   pointer too.

   Call ftrace_graph_init() on the task first which will NULL out the
   new tasks's ret_stack and if the copy fails, it will not free
   anything.

 - Remove FGRAPH_MAX_INDEX

   The macro FGRAPH_MAX_INDEX was added but never used. Remove it.

 - Save ent_size in function graph printing of nested functions

   The function graph tracer needs to look at the next event to see if
   the next event is the return of the current function entry. If it is,
   it prints a single line:

	ktime_get();

   Otherwise it prints it like a nested function:

	tick_nohz_irq_exit() {
	    ktime_get();
	    kcpustat_irq_exit();
	}

   In order to look at the next event, it must save the current event so
   that it has the information to print from it. It saves the event in
   the iterator descriptor called "ent". What it doesn't save is the
   ent_size of the event which is now used to know if the function graph
   arguments are to be printed. The peek doesn't save the size so the
   size used happens to be that of the size of the last event that was
   seen.

   Save the entry event size in the iterator descriptor so that the
   correct size is used.

 - Fix several errors with freeing data in the histogram code

   The histogram code had a lot of leaked or or incorrect accounting
   when failures happen. Correct them.

 - Fix histogram regression of .percent and .graph modifiers

   Up until 6.3 histogram values could have "percent" or "graph"
   modifiers that changed how they were printed. But a change that added
   restricting histograms values from being strings, stack traces and
   other modifiers inadvertently prevented them from using the percent
   and graph modifiers, which were legal use cases for values.

   Put back the percent and graph modifiers.

 - Fix various typos in the comments

 - Set the trace_clock before initializing a histogram with clock
   argument

   The histogram API allows the user to specific which trace clock to
   use via a "clock=" string. The histogram is set up first before the
   clock is checked. If the passed in clock is not valid, it exits
   without fully fixing up the histogram leaving it on the list and a
   use-after-free can trigger.

   Update the clock argument first and if it fails then exit gracefully
   before the histogram trigger is placed on any lists.

 - Restore :mod: trailer after parsing in ftrace_set_clr_event

   The function ftrace_set_clr_event() modifies the parse string and
   needs to put it back to what was passed in. It searches for ":mod:"
   via a strsep() but fails to put back the first ':' in the string.

   Add back the ':' in the passed in string.

 - Take trace_array reference when opening a tracer options file

   The options files are dynamically created and some tracers add their
   own options. When a tracer adds their own list of options, the
   trace_array holding them has an array to hold the list of options for
   each tracer. This array increases in size via a krealloc(), and the
   new entry gets a newly allocated array to hold the options of the new
   tracer being added.

   The element in each entry of the tracer's option array holds a
   pointer back to the trace_array, a pointer to the tracer it is
   associated to, a pointer to the flags of the option.

   The issue is that these arrays are freed when the trace_array is
   freed when its instance it represents is removed from the instances
   directory. There's a race that an open of one of these options files
   can happen when the instance is being removed.

   Add a new helper function to be called by the open function of the
   options file to iterate all existing trace_arrays under a lock and
   find the one that has the given option element in one of it's tracer
   arrays. If found, then update the associated trace_array's reference
   counter to keep it from being freed. If not found, have the open call
   return -ENODEV.

 - Disable interrupts when acquiring the lock in rb_wake_up_waiters()

   The function rb_wake_up_waiters() assumes it will be called in
   interrupt context and does not disable irqs when taking
   cpu_buffer->reader_lock, which can be called in hard interrupt
   context. The issue is in PREEMPT_RT, this function is called in
   thread context leaving this lock open to a deadlock.

   Take the lock with interrupts disabled.

 - Use rcu_assign_pointer() for tmp_ops filter hash

   The tmp_ops used in update_ftrace_direct_mod() assigns its
   filter_hash field directly, but that field is annotated as __rcu and
   sparse complains. Assign it with rcu_assign_pointer()

 - Fix use-after-free in enable_trigger_private_data_free()

   The trace_event_call is accessed through the event_trigger_data's
   trace_event_file pointer to put the trace_event_call on freeing. The
   issue is that the trace_event_file data may have been freed already
   causing a use-after-free. Add a field to the event_trigger_data that
   points directly to the trace_event_call so that it can decrement its
   reference directly without needing to go through the
   trace_event_file.

 - Fix accounting of buffer data remote headers

   trace_buffer_desc_size() and trace_remote_alloc_buffer() undercount
   the number of pages is needed for the asked for size as it doesn't
   take into account the meta data on each page. Add a helper function
   to do the calculation properly and use that in these functions.

 - Catch nr_page_va overflow in ring_buffer_desc sizing

   The number of pages per remote ring buffer is capped by
   ring_buffer_desc::nr_page_va (32 bits). A buffer_size large enough to
   overflow that field would silently allocate a descriptor smaller than
   what was asked for.

 - Do not resize the subbuf order if any per_cpu buffer is disabled

   The mmapping of ring buffers disables resizing the subbuffers, but it
   is done per-cpu whereas the subbuf size change is done for all the
   per_cpu buffers under the buffer->mutex. It could change the size of
   some while the mapping is happening on others. Have the resize of the
   subbuf order check all the per_cpu buffers under the lock to see if
   any of them is disabled before starting and causing an inconsistency
   between buffers that are being mapped.

* tag 'trace-v7.3-rc2' of git://git.kernel.org/pub/scm/linux/kernel/git/trace/linux-trace: (25 commits)
  ring-buffer: Check resize_disabled before publishing the new subbuf order
  tracing/remotes: Catch nr_page_va overflow in ring_buffer_desc sizing
  tracing/remotes: Account for ring buffer page header in size calculation
  tracing: Don't dereference trace_event_file in deferred trigger free
  ftrace: Use rcu_assign_pointer() for tmp_ops filter hash
  ring-buffer: Acquire the lock with irqsave in rb_wake_up_waiters()
  tracing: Take trace_array reference when opening a tracer options file
  tracing: Fix ring_buffer_read_page_size() kernel-doc
  tracing: Restore :mod: trailer after parsing in ftrace_set_clr_event()
  tracing: Fix memory corruption from a "STACKTRACE" histogram key
  tracing: Fix memory corruption from the histogram stacktrace modifier
  tracing: Undo the registration when enabling the histogram trigger fails
  tracing: Take the reference before publishing the named histogram trigger
  tracing: Set the trace clock before registering the histogram trigger
  tracing: Fix typo "preceeded" in comment
  tracing: Fix typo "availabe" in comment
  tracing: Let histogram values keep the percent and graph modifiers
  tracing: Keep the entry count when the histogram stats allocation fails
  tracing: Free histogram the field rejected for a bad modifier
  tracing: Free histogram the var ref when its initialization fails
  ...

1 files changed, 888 insertions, 0 deletions

diff --git a/tools/perf/tests/code-reading.c b/tools/perf/tests/code-reading.c
new file mode 100644
index 000000000..f0e8ea875
--- /dev/null
+++ b/tools/perf/tests/code-reading.c
@@ -0,0 +1,888 @@
+// SPDX-License-Identifier: GPL-2.0
+#include <errno.h>
+#include <linux/kconfig.h>
+#include <linux/kernel.h>
+#include <linux/rbtree.h>
+#include <linux/types.h>
+#include <linux/zalloc.h>
+#include <inttypes.h>
+#include <stdlib.h>
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <sys/param.h>
+#include <sys/utsname.h>
+#include <perf/cpumap.h>
+#include <perf/evlist.h>
+#include <perf/mmap.h>
+
+#include "debug.h"
+#include "dso.h"
+#include "env.h"
+#include "parse-events.h"
+#include "evlist.h"
+#include "evsel.h"
+#include "thread_map.h"
+#include "machine.h"
+#include "map.h"
+#include "symbol.h"
+#include "event.h"
+#include "record.h"
+#include "util/mmap.h"
+#include "util/string2.h"
+#include "util/synthetic-events.h"
+#include "util/util.h"
+#include "thread.h"
+
+#include "tests.h"
+
+#include <linux/ctype.h>
+
+#define BUFSZ	1024
+#define READLEN	128
+
+struct tested_section {
+	struct rb_node rb_node;
+	u64 addr;
+	char *path;
+};
+
+static bool tested_code_insert_or_exists(const char *path, u64 addr,
+					 struct rb_root *tested_sections)
+{
+	struct rb_node **node = &tested_sections->rb_node;
+	struct rb_node *parent = NULL;
+	struct tested_section *data;
+
+	while (*node) {
+		int cmp;
+
+		parent = *node;
+		data = rb_entry(*node, struct tested_section, rb_node);
+		cmp = strcmp(path, data->path);
+		if (!cmp) {
+			if (addr < data->addr)
+				cmp = -1;
+			else if (addr > data->addr)
+				cmp = 1;
+			else
+				return true; /* already tested */
+		}
+
+		if (cmp < 0)
+			node = &(*node)->rb_left;
+		else
+			node = &(*node)->rb_right;
+	}
+
+	data = zalloc(sizeof(*data));
+	if (!data)
+		return true;
+
+	data->addr = addr;
+	data->path = strdup(path);
+	if (!data->path) {
+		free(data);
+		return true;
+	}
+	rb_link_node(&data->rb_node, parent, node);
+	rb_insert_color(&data->rb_node, tested_sections);
+	return false;
+}
+
+static void tested_sections__free(struct rb_root *root)
+{
+	while (!RB_EMPTY_ROOT(root)) {
+		struct rb_node *node = rb_first(root);
+		struct tested_section *ts = rb_entry(node,
+						     struct tested_section,
+						     rb_node);
+
+		rb_erase(node, root);
+		free(ts->path);
+		free(ts);
+	}
+}
+
+static size_t read_objdump_chunk(const char **line, unsigned char **buf,
+				 size_t *buf_len)
+{
+	size_t bytes_read = 0;
+	unsigned char *chunk_start = *buf;
+
+	/* Read bytes */
+	while (*buf_len > 0) {
+		char c1, c2;
+
+		/* Get 2 hex digits */
+		c1 = *(*line)++;
+		if (!isxdigit(c1))
+			break;
+		c2 = *(*line)++;
+		if (!isxdigit(c2))
+			break;
+
+		/* Store byte and advance buf */
+		**buf = (hex(c1) << 4) | hex(c2);
+		(*buf)++;
+		(*buf_len)--;
+		bytes_read++;
+
+		/* End of chunk? */
+		if (isspace(**line))
+			break;
+	}
+
+	/*
+	 * objdump will display raw insn as LE if code endian
+	 * is LE and bytes_per_chunk > 1. In that case reverse
+	 * the chunk we just read.
+	 *
+	 * see disassemble_bytes() at binutils/objdump.c for details
+	 * how objdump chooses display endian)
+	 */
+	if (bytes_read > 1 && !host_is_bigendian()) {
+		unsigned char *chunk_end = chunk_start + bytes_read - 1;
+		unsigned char tmp;
+
+		while (chunk_start < chunk_end) {
+			tmp = *chunk_start;
+			*chunk_start = *chunk_end;
+			*chunk_end = tmp;
+			chunk_start++;
+			chunk_end--;
+		}
+	}
+
+	return bytes_read;
+}
+
+static size_t read_objdump_line(const char *line, unsigned char *buf,
+				size_t buf_len)
+{
+	const char *p;
+	size_t ret, bytes_read = 0;
+
+	/* Skip to a colon */
+	p = strchr(line, ':');
+	if (!p)
+		return 0;
+	p++;
+
+	/* Skip initial spaces */
+	while (*p) {
+		if (!isspace(*p))
+			break;
+		p++;
+	}
+
+	do {
+		ret = read_objdump_chunk(&p, &buf, &buf_len);
+		bytes_read += ret;
+		p++;
+	} while (ret > 0);
+
+	/* return number of successfully read bytes */
+	return bytes_read;
+}
+
+static int read_objdump_output(FILE *f, void *buf, size_t *len, u64 start_addr)
+{
+	char *line = NULL;
+	size_t line_len, off_last = 0;
+	ssize_t ret;
+	int err = 0;
+	u64 addr, last_addr = start_addr;
+
+	while (off_last < *len) {
+		size_t off, read_bytes, written_bytes;
+		unsigned char tmp[BUFSZ];
+
+		ret = getline(&line, &line_len, f);
+		if (feof(f))
+			break;
+		if (ret < 0) {
+			pr_debug("getline failed\n");
+			err = -1;
+			break;
+		}
+
+		/* read objdump data into temporary buffer */
+		read_bytes = read_objdump_line(line, tmp, sizeof(tmp));
+		if (!read_bytes)
+			continue;
+
+		if (sscanf(line, "%"PRIx64, &addr) != 1)
+			continue;
+		if (addr < last_addr) {
+			pr_debug("addr going backwards, read beyond section?\n");
+			break;
+		}
+		last_addr = addr;
+
+		/* copy it from temporary buffer to 'buf' according
+		 * to address on current objdump line */
+		off = addr - start_addr;
+		if (off >= *len)
+			break;
+		written_bytes = MIN(read_bytes, *len - off);
+		memcpy(buf + off, tmp, written_bytes);
+		off_last = off + written_bytes;
+	}
+
+	/* len returns number of bytes that could not be read */
+	*len -= off_last;
+
+	free(line);
+
+	return err;
+}
+
+/*
+ * Only gets GNU objdump version. Returns 0 for llvm-objdump.
+ */
+static int objdump_version(void)
+{
+	size_t line_len;
+	char cmd[PATH_MAX * 2];
+	char *line = NULL;
+	const char *fmt;
+	FILE *f;
+	int ret;
+
+	int version_tmp, version_num = 0;
+	char *version = 0, *token;
+
+	fmt = "%s --version";
+	ret = snprintf(cmd, sizeof(cmd), fmt, test_objdump_path);
+	if (ret <= 0 || (size_t)ret >= sizeof(cmd))
+		return -1;
+	/* Ignore objdump errors */
+	strcat(cmd, " 2>/dev/null");
+	f = popen(cmd, "r");
+	if (!f) {
+		pr_debug("popen failed\n");
+		return -1;
+	}
+	/* Get first line of objdump --version output */
+	ret = getline(&line, &line_len, f);
+	pclose(f);
+	if (ret < 0) {
+		pr_debug("getline failed\n");
+		return -1;
+	}
+
+	token = strsep(&line, " ");
+	if (token != NULL && !strcmp(token, "GNU")) {
+		// version is last part of first line of objdump --version output.
+		while ((token = strsep(&line, " ")))
+			version = token;
+
+		// Convert version into a format we can compare with
+		token = strsep(&version, ".");
+		version_num = atoi(token);
+		if (version_num)
+			version_num *= 10000;
+
+		token = strsep(&version, ".");
+		version_tmp = atoi(token);
+		if (token)
+			version_num += version_tmp * 100;
+
+		token = strsep(&version, ".");
+		version_tmp = atoi(token);
+		if (token)
+			version_num += version_tmp;
+	}
+
+	return version_num;
+}
+
+static int read_via_objdump(const char *filename, u64 addr, void *buf,
+			    size_t len)
+{
+	u64 stop_address = addr + len;
+	struct utsname uname_buf;
+	char cmd[PATH_MAX * 2];
+	const char *fmt;
+	FILE *f;
+	int ret;
+
+	ret = uname(&uname_buf);
+	if (ret) {
+		pr_debug("uname failed\n");
+		return -1;
+	}
+
+	if (!strncmp(uname_buf.machine, "riscv", 5)) {
+		int version = objdump_version();
+
+		/* Default to this workaround if version parsing fails */
+		if (version < 0 || version > 24100) {
+			/*
+			 * Starting at riscv objdump version 2.41, dumping in
+			 * the middle of an instruction is not supported. riscv
+			 * instructions are aligned along 2-byte intervals and
+			 * can be either 2-bytes or 4-bytes. This makes it
+			 * possible that the stop-address lands in the middle of
+			 * a 4-byte instruction. Increase the stop_address by
+			 * two to ensure an instruction is not cut in half, but
+			 * leave the len as-is so only the expected number of
+			 * bytes are collected.
+			 */
+			stop_address += 2;
+		}
+	}
+
+	fmt = "%s -z -d --start-address=0x%"PRIx64" --stop-address=0x%"PRIx64" %s";
+	ret = snprintf(cmd, sizeof(cmd), fmt, test_objdump_path, addr, stop_address,
+		       filename);
+	if (ret <= 0 || (size_t)ret >= sizeof(cmd))
+		return -1;
+
+	pr_debug("Objdump command is: %s\n", cmd);
+
+	/* Ignore objdump errors */
+	strcat(cmd, " 2>/dev/null");
+
+	f = popen(cmd, "r");
+	if (!f) {
+		pr_debug("popen failed\n");
+		return -1;
+	}
+
+	ret = read_objdump_output(f, buf, &len, addr);
+	if (len) {
+		pr_debug("objdump read too few bytes: %zd\n", len);
+		if (!ret)
+			ret = len;
+	}
+
+	pclose(f);
+
+	return ret;
+}
+
+static void dump_buf(unsigned char *buf, size_t len)
+{
+	size_t i;
+
+	for (i = 0; i < len; i++) {
+		pr_debug("0x%02x ", buf[i]);
+		if (i % 16 == 15)
+			pr_debug("\n");
+	}
+	pr_debug("\n");
+}
+
+static int read_object_code(u64 addr, size_t len, u8 cpumode,
+			    struct thread *thread,
+			    struct rb_root *tested_sections)
+{
+	struct addr_location al;
+	unsigned char buf1[BUFSZ] = {0};
+	unsigned char buf2[BUFSZ] = {0};
+	size_t ret_len;
+	u64 objdump_addr;
+	u64 skip_addr;
+	const char *objdump_name;
+	char decomp_name[KMOD_DECOMP_LEN];
+	bool decomp = false;
+	int ret, err = 0;
+	struct dso *dso;
+
+	pr_debug("Reading object code for memory address: %#"PRIx64"\n", addr);
+
+	addr_location__init(&al);
+	if (!thread__find_map(thread, cpumode, addr, &al) || !map__dso(al.map)) {
+		if (cpumode == PERF_RECORD_MISC_HYPERVISOR) {
+			pr_debug("Hypervisor address can not be resolved - skipping\n");
+			goto out;
+		}
+
+		pr_debug("thread__find_map failed\n");
+		err = -1;
+		goto out;
+	}
+	dso = map__dso(al.map);
+	pr_debug("File is: %s\n", dso__long_name(dso));
+
+	if (dso__symtab_type(dso) == DSO_BINARY_TYPE__KALLSYMS && !dso__is_kcore(dso)) {
+		pr_debug("Unexpected kernel address - skipping\n");
+		goto out;
+	}
+
+	/*
+	 * Don't retest the same addresses. objdump struggles with kcore - try
+	 * each map only once even if the address is different.
+	 */
+	skip_addr = dso__is_kcore(dso) ? map__start(al.map) : al.addr;
+	if (tested_code_insert_or_exists(dso__long_name(dso), skip_addr,
+					 tested_sections)) {
+		pr_debug("Already tested %s @ %#"PRIx64" - skipping\n",
+			 dso__long_name(dso), skip_addr);
+		goto out;
+	}
+
+	pr_debug("On file address is: %#"PRIx64"\n", al.addr);
+
+	if (len > BUFSZ)
+		len = BUFSZ;
+
+	/* Do not go off the map */
+	if (addr + len > map__end(al.map))
+		len = map__end(al.map) - addr;
+
+	/*
+	 * Some architectures (ex: powerpc) have stubs (trampolines) in kernel
+	 * modules to manage long jumps. Check if the ip offset falls in stubs
+	 * sections for kernel modules. And skip module address after text end
+	 */
+	if (dso__is_kmod(dso) && al.addr > dso__text_end(dso)) {
+		pr_debug("skipping the module address %#"PRIx64" after text end\n", al.addr);
+		goto out;
+	}
+
+	/* Read the object code using perf */
+	ret_len = dso__data_read_offset(dso, maps__machine(thread__maps(thread)),
+					al.addr, buf1, len);
+	if (ret_len != len) {
+		pr_debug("dso__data_read_offset failed\n");
+		err = -1;
+		goto out;
+	}
+
+	/*
+	 * Converting addresses for use by objdump requires more information.
+	 * map__load() does that.  See map__rip_2objdump() for details.
+	 */
+	if (map__load(al.map)) {
+		err = -1;
+		goto out;
+	}
+
+	objdump_name = dso__long_name(dso);
+	if (dso__needs_decompress(dso)) {
+		if (dso__decompress_kmodule_path(dso, objdump_name,
+						 decomp_name,
+						 sizeof(decomp_name)) < 0) {
+			pr_debug("decompression failed\n");
+			err = -1;
+			goto out;
+		}
+
+		/* empty pathname means file wasn't actually compressed */
+		if (decomp_name[0] != '\0') {
+			decomp = true;
+			objdump_name = decomp_name;
+		}
+	}
+
+	/* Read the object code using objdump */
+	objdump_addr = map__rip_2objdump(al.map, al.addr);
+	ret = read_via_objdump(objdump_name, objdump_addr, buf2, len);
+
+	if (decomp)
+		unlink(objdump_name);
+
+	if (ret > 0) {
+		/*
+		 * The kernel maps are inaccurate - assume objdump is right in
+		 * that case.
+		 */
+		if (cpumode == PERF_RECORD_MISC_KERNEL ||
+		    cpumode == PERF_RECORD_MISC_GUEST_KERNEL) {
+			len -= ret;
+			if (len) {
+				pr_debug("Reducing len to %zu\n", len);
+			} else if (dso__is_kcore(dso)) {
+				/*
+				 * objdump cannot handle very large segments
+				 * that may be found in kcore.
+				 */
+				pr_debug("objdump failed for kcore");
+				pr_debug(" - skipping\n");
+			} else {
+				err = -1;
+			}
+			goto out;
+		}
+	}
+	if (ret < 0) {
+		pr_debug("read_via_objdump failed\n");
+		err = -1;
+		goto out;
+	}
+
+	/* The results should be identical */
+	if (memcmp(buf1, buf2, len)) {
+		pr_debug("Bytes read differ from those read by objdump\n");
+		pr_debug("buf1 (dso):\n");
+		dump_buf(buf1, len);
+		pr_debug("buf2 (objdump):\n");
+		dump_buf(buf2, len);
+		err = -1;
+		goto out;
+	}
+	pr_debug("Bytes read match those read by objdump\n");
+out:
+	addr_location__exit(&al);
+	return err;
+}
+
+static int process_sample_event(struct machine *machine, struct evlist *evlist,
+				union perf_event *event,
+				struct rb_root *tested_sections)
+{
+	struct perf_sample sample;
+	struct thread *thread;
+	int ret;
+
+	perf_sample__init(&sample, /*all=*/false);
+	ret = evlist__parse_sample(evlist, event, &sample);
+	if (ret) {
+		pr_debug("evlist__parse_sample failed\n");
+		ret = -1;
+		goto out;
+	}
+
+	thread = machine__findnew_thread(machine, sample.pid, sample.tid);
+	if (!thread) {
+		pr_debug("machine__findnew_thread failed\n");
+		ret = -1;
+		goto out;
+	}
+
+	ret = read_object_code(sample.ip, READLEN, sample.cpumode, thread,
+			       tested_sections);
+	thread__put(thread);
+out:
+	perf_sample__exit(&sample);
+	return ret;
+}
+
+static int process_event(struct machine *machine, struct evlist *evlist,
+			 union perf_event *event, struct rb_root *tested_sections)
+{
+	if (event->header.type == PERF_RECORD_SAMPLE)
+		return process_sample_event(machine, evlist, event,
+					    tested_sections);
+
+	if (event->header.type == PERF_RECORD_THROTTLE ||
+	    event->header.type == PERF_RECORD_UNTHROTTLE)
+		return 0;
+
+	if (event->header.type < PERF_RECORD_MAX) {
+		int ret;
+
+		ret = machine__process_event(machine, event, NULL);
+		if (ret < 0)
+			pr_debug("machine__process_event failed, event type %u\n",
+				 event->header.type);
+		return ret;
+	}
+
+	return 0;
+}
+
+static int process_events(struct machine *machine, struct evlist *evlist,
+			  struct rb_root *tested_sections)
+{
+	union perf_event *event;
+	struct mmap *md;
+	int i, ret;
+
+	for (i = 0; i < evlist__core(evlist)->nr_mmaps; i++) {
+		md = &evlist__mmap(evlist)[i];
+		if (perf_mmap__read_init(&md->core) < 0)
+			continue;
+
+		while ((event = perf_mmap__read_event(&md->core)) != NULL) {
+			ret = process_event(machine, evlist, event, tested_sections);
+			perf_mmap__consume(&md->core);
+			if (ret < 0)
+				return ret;
+		}
+		perf_mmap__read_done(&md->core);
+	}
+	return 0;
+}
+
+static int comp(const void *a, const void *b)
+{
+	return *(int *)a - *(int *)b;
+}
+
+static void do_sort_something(void)
+{
+	int buf[40960], i;
+
+	for (i = 0; i < (int)ARRAY_SIZE(buf); i++)
+		buf[i] = ARRAY_SIZE(buf) - i - 1;
+
+	qsort(buf, ARRAY_SIZE(buf), sizeof(int), comp);
+
+	for (i = 0; i < (int)ARRAY_SIZE(buf); i++) {
+		if (buf[i] != i) {
+			pr_debug("qsort failed\n");
+			break;
+		}
+	}
+}
+
+static void sort_something(void)
+{
+	int i;
+
+	for (i = 0; i < 10; i++)
+		do_sort_something();
+}
+
+static void syscall_something(void)
+{
+	int pipefd[2];
+	int i;
+
+	for (i = 0; i < 1000; i++) {
+		if (pipe(pipefd) < 0) {
+			pr_debug("pipe failed\n");
+			break;
+		}
+		close(pipefd[1]);
+		close(pipefd[0]);
+	}
+}
+
+static void fs_something(void)
+{
+	const char *test_file_name = "temp-perf-code-reading-test-file--";
+	FILE *f;
+	int i;
+
+	for (i = 0; i < 1000; i++) {
+		f = fopen(test_file_name, "w+");
+		if (f) {
+			fclose(f);
+			unlink(test_file_name);
+		}
+	}
+}
+
+static void do_something(void)
+{
+	fs_something();
+
+	sort_something();
+
+	syscall_something();
+}
+
+enum {
+	TEST_CODE_READING_OK,
+	TEST_CODE_READING_NO_VMLINUX,
+	TEST_CODE_READING_NO_KCORE,
+	TEST_CODE_READING_NO_ACCESS,
+	TEST_CODE_READING_NO_KERNEL_OBJ,
+};
+
+static int do_test_code_reading(bool try_kcore)
+{
+	struct machine *machine;
+	struct thread *thread;
+	struct record_opts opts = {
+		.mmap_pages	     = UINT_MAX,
+		.user_freq	     = UINT_MAX,
+		.user_interval	     = ULLONG_MAX,
+		.freq		     = 500,
+		.target		     = {
+			.uses_mmap   = true,
+		},
+	};
+	struct rb_root tested_sections = RB_ROOT;
+	struct perf_thread_map *threads = NULL;
+	struct perf_cpu_map *cpus = NULL;
+	struct evlist *evlist = NULL;
+	struct evsel *evsel = NULL;
+	int err = -1, ret;
+	pid_t pid;
+	struct map *map;
+	bool have_vmlinux, have_kcore;
+	struct dso *dso;
+	const char *events[] = { "cpu-cycles", "cpu-cycles:u", "cpu-clock", "cpu-clock:u", NULL };
+	int evidx = 0;
+	struct perf_env host_env;
+
+	pid = getpid();
+
+	perf_env__init(&host_env);
+	machine = machine__new_host(&host_env);
+
+	ret = machine__create_kernel_maps(machine);
+	if (ret < 0) {
+		pr_debug("machine__create_kernel_maps failed\n");
+		goto out_err;
+	}
+
+	/* Force the use of kallsyms instead of vmlinux to try kcore */
+	if (try_kcore)
+		symbol_conf.kallsyms_name = "/proc/kallsyms";
+
+	/* Load kernel map */
+	map = machine__kernel_map(machine);
+	ret = map__load(map);
+	if (ret < 0) {
+		pr_debug("map__load failed\n");
+		goto out_err;
+	}
+	dso = map__dso(map);
+	have_vmlinux = dso__is_vmlinux(dso);
+	have_kcore = dso__is_kcore(dso);
+
+	/* 2nd time through we just try kcore */
+	if (try_kcore && !have_kcore)
+		return TEST_CODE_READING_NO_KCORE;
+
+	/* No point getting kernel events if there is no kernel object */
+	if (!have_vmlinux && !have_kcore)
+		evidx++;
+
+	threads = thread_map__new_by_tid(pid);
+	if (!threads) {
+		pr_debug("thread_map__new_by_tid failed\n");
+		goto out_err;
+	}
+
+	ret = perf_event__synthesize_thread_map(NULL, threads,
+						perf_event__process, machine,
+						true, false);
+	if (ret < 0) {
+		pr_debug("perf_event__synthesize_thread_map failed\n");
+		goto out_err;
+	}
+
+	thread = machine__findnew_thread(machine, pid, pid);
+	if (!thread) {
+		pr_debug("machine__findnew_thread failed\n");
+		goto out_put;
+	}
+
+	cpus = perf_cpu_map__new_online_cpus();
+	if (!cpus) {
+		pr_debug("perf_cpu_map__new failed\n");
+		goto out_put;
+	}
+
+	while (events[evidx]) {
+		const char *str;
+
+		evlist = evlist__new();
+		if (!evlist) {
+			pr_debug("evlist__new failed\n");
+			goto out_put;
+		}
+
+		perf_evlist__set_maps(evlist__core(evlist), cpus, threads);
+
+		str = events[evidx];
+		pr_debug("Parsing event '%s'\n", str);
+		ret = parse_event(evlist, str);
+		if (ret < 0) {
+			pr_debug("parse_events failed\n");
+			goto out_put;
+		}
+
+		evlist__config(evlist, &opts, NULL);
+
+		evlist__for_each_entry(evlist, evsel) {
+			evsel->core.attr.comm = 1;
+			evsel->core.attr.disabled = 1;
+			evsel->core.attr.enable_on_exec = 0;
+		}
+
+		ret = evlist__open(evlist);
+		if (ret < 0) {
+			evidx++;
+
+			if (events[evidx] == NULL && verbose > 0) {
+				char errbuf[512];
+				evlist__strerror_open(evlist, errno, errbuf, sizeof(errbuf));
+				pr_debug("perf_evlist__open() failed!\n%s\n", errbuf);
+			}
+
+			perf_evlist__set_maps(evlist__core(evlist), NULL, NULL);
+			evlist__put(evlist);
+			evlist = NULL;
+			continue;
+		}
+		break;
+	}
+
+	if (events[evidx] == NULL)
+		goto out_put;
+
+	ret = evlist__do_mmap(evlist, UINT_MAX);
+	if (ret < 0) {
+		pr_debug("evlist__mmap failed\n");
+		goto out_put;
+	}
+
+	evlist__enable(evlist);
+
+	do_something();
+
+	evlist__disable(evlist);
+
+	ret = process_events(machine, evlist, &tested_sections);
+	if (ret < 0)
+		goto out_put;
+
+	if (!have_vmlinux && !have_kcore && !try_kcore)
+		err = TEST_CODE_READING_NO_KERNEL_OBJ;
+	else if (!have_vmlinux && !try_kcore)
+		err = TEST_CODE_READING_NO_VMLINUX;
+	else if (strstr(events[evidx], ":u"))
+		err = TEST_CODE_READING_NO_ACCESS;
+	else
+		err = TEST_CODE_READING_OK;
+out_put:
+	thread__put(thread);
+out_err:
+	evlist__put(evlist);
+	perf_cpu_map__put(cpus);
+	perf_thread_map__put(threads);
+	machine__delete(machine);
+	perf_env__exit(&host_env);
+	tested_sections__free(&tested_sections);
+
+	return err;
+}
+
+static int test__code_reading(struct test_suite *test __maybe_unused, int subtest __maybe_unused)
+{
+	int ret;
+
+	ret = do_test_code_reading(false);
+	if (!ret)
+		ret = do_test_code_reading(true);
+
+	switch (ret) {
+	case TEST_CODE_READING_OK:
+		return 0;
+	case TEST_CODE_READING_NO_VMLINUX:
+		pr_debug("no vmlinux\n");
+		return 0;
+	case TEST_CODE_READING_NO_KCORE:
+		pr_debug("no kcore\n");
+		return 0;
+	case TEST_CODE_READING_NO_ACCESS:
+		pr_debug("no access\n");
+		return 0;
+	case TEST_CODE_READING_NO_KERNEL_OBJ:
+		pr_debug("no kernel obj\n");
+		return 0;
+	default:
+		return -1;
+	};
+}
+
+DEFINE_SUITE("Object code reading", code_reading);