Commit b5906f5f authored by SeongJae Park's avatar SeongJae Park Committed by Andrew Morton
Browse files

selftests/damon: add a test for update_schemes_tried_regions sysfs command

Add a selftest for verifying the accuracy of DAMON's access monitoring
functionality.  The test starts a program of artificial access pattern,
monitor the access pattern using DAMON, and check if DAMON finds expected
amount of hot data region (working set size) with only acceptable error
rate.

Note that the acceptable error rate is set with only naive assumptions and
small number of tests.  Hence failures of the test may not always mean
DAMON is broken.  Rather than that, those could be a signal to better
understand the real accuracy level of DAMON in wider environments.  Based
on further finding, we could optimize DAMON or adjust the expectation of
the test.

Link: https://lkml.kernel.org/r/20231212194810.54457-5-sj@kernel.org


Signed-off-by: default avatarSeongJae Park <sj@kernel.org>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: default avatarAndrew Morton <akpm@linux-foundation.org>
parent 3402c6ce
Loading
Loading
Loading
Loading
+2 −0
Original line number Diff line number Diff line
@@ -2,6 +2,7 @@
# Makefile for damon selftests

TEST_GEN_FILES += huge_count_read_write
TEST_GEN_FILES += access_memory

TEST_FILES = _chk_dependency.sh _debugfs_common.sh
TEST_PROGS = debugfs_attrs.sh debugfs_schemes.sh debugfs_target_ids.sh
@@ -9,6 +10,7 @@ TEST_PROGS += debugfs_empty_targets.sh debugfs_huge_count_read_write.sh
TEST_PROGS += debugfs_duplicate_context_creation.sh
TEST_PROGS += debugfs_rm_non_contexts.sh
TEST_PROGS += sysfs.sh sysfs_update_removed_scheme_dir.sh
TEST_PROGS += sysfs_update_schemes_tried_regions_wss_estimation.py
TEST_PROGS += reclaim.sh lru_sort.sh

include ../lib.mk
+41 −0
Original line number Diff line number Diff line
// SPDX-License-Identifier: GPL-2.0
/*
 * Artificial memory access program for testing DAMON.
 */

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <time.h>

int main(int argc, char *argv[])
{
	char **regions;
	clock_t start_clock;
	int nr_regions;
	int sz_region;
	int access_time_ms;
	int i;

	if (argc != 4) {
		printf("Usage: %s <number> <size (bytes)> <time (ms)>\n",
				argv[0]);
		return -1;
	}

	nr_regions = atoi(argv[1]);
	sz_region = atoi(argv[2]);
	access_time_ms = atoi(argv[3]);

	regions = malloc(sizeof(*regions) * nr_regions);
	for (i = 0; i < nr_regions; i++)
		regions[i] = malloc(sz_region);

	for (i = 0; i < nr_regions; i++) {
		start_clock = clock();
		while ((clock() - start_clock) * 1000 / CLOCKS_PER_SEC <
				access_time_ms)
			memset(regions[i], i, 1024 * 1024 * 10);
	}
	return 0;
}
+55 −0
Original line number Diff line number Diff line
#!/usr/bin/env python3
# SPDX-License-Identifier: GPL-2.0

import subprocess
import time

import _damon_sysfs

def main():
    # access two 10 MiB memory regions, 2 second per each
    sz_region = 10 * 1024 * 1024
    proc = subprocess.Popen(['./access_memory', '2', '%d' % sz_region, '2000'])
    kdamonds = _damon_sysfs.Kdamonds([_damon_sysfs.Kdamond(
            contexts=[_damon_sysfs.DamonCtx(
                ops='vaddr',
                targets=[_damon_sysfs.DamonTarget(pid=proc.pid)],
                schemes=[_damon_sysfs.Damos(
                    access_pattern=_damon_sysfs.DamosAccessPattern(
                        # >= 25% access rate, >= 200ms age
                        nr_accesses=[5, 20], age=[2, 2**64 - 1]))] # schemes
                )] # contexts
            )]) # kdamonds

    err = kdamonds.start()
    if err != None:
        print('kdmaond start failed: %s' % err)
        exit(1)

    wss_collected = []
    while proc.poll() == None:
        time.sleep(0.1)
        err = kdamonds.kdamonds[0].update_schemes_tried_bytes()
        if err != None:
            print('tried bytes update failed: %s' % err)
            exit(1)

        wss_collected.append(
                kdamonds.kdamonds[0].contexts[0].schemes[0].tried_bytes)

    wss_collected.sort()
    acceptable_error_rate = 0.2
    for percentile in [50, 75]:
        sample = wss_collected[int(len(wss_collected) * percentile / 100)]
        error_rate = abs(sample - sz_region) / sz_region
        print('%d-th percentile (%d) error %f' %
                (percentile, sample, error_rate))
        if error_rate > acceptable_error_rate:
            print('the error rate is not acceptable (> %f)' %
                    acceptable_error_rate)
            print('samples are as below')
            print('\n'.join(['%d' % wss for wss in wss_collected]))
            exit(1)

if __name__ == '__main__':
    main()