cgroup: test for --manage-cgroups=ignore

Test to ensure that --manage-cgroups=ignore works correctly.

Signed-off-by: Adrian Reber <areber@redhat.com>

test/jenkins/crit.sh

@@ -2,6 +2,6 @@
 set -e
 source `dirname $0`/criu-lib.sh
 prep
-./test/zdtm.py run --all -f best -x maps04 -x cgroup02 --norst --keep-img always || fail
+./test/zdtm.py run --all -f best -x maps04 -x cgroup02 -x cgroup_ignore --norst --keep-img always || fail
 PYTHONPATH="$(pwd)/lib/" ./test/crit-recode.py || fail
 exit 0

test/jenkins/criu-dump.sh

@@ -5,4 +5,4 @@ set -e
 source `dirname $0`/criu-lib.sh
 prep
 mount_tmpfs_to_dump
-./test/zdtm.py run --all --keep-going --report report --parallel 4 --norst -x 'maps04' -x 'cgroup02' || fail
+./test/zdtm.py run --all --keep-going --report report --parallel 4 --norst -x 'maps04' -x 'cgroup02' -x 'cgroup_ignore' || fail

test/zdtm/lib/zdtmtst.h

@@ -176,4 +176,30 @@ extern dev_t get_mapping_dev(void *addr);
 		___ret;                                              \
 	})
 
+#ifndef TEMP_FAILURE_RETRY
+#define TEMP_FAILURE_RETRY(expression)                     \
+	(__extension__({                                   \
+		long int __result;                         \
+		do                                         \
+			__result = (long int)(expression); \
+		while (__result < 0 && errno == EINTR);    \
+		__result;                                  \
+	}))
+#endif
+
+#define cleanup_close __attribute__((cleanup(cleanup_closep)))
+#define cleanup_free  __attribute__((cleanup(cleanup_freep)))
+static inline void cleanup_freep(void *p)
+{
+	void **pp = (void **)p;
+	free(*pp);
+}
+
+static inline void cleanup_closep(void *p)
+{
+	int *pp = (int *)p;
+	if (*pp >= 0)
+		TEMP_FAILURE_RETRY(close(*pp));
+}
+
 #endif /* _VIMITESU_H_ */

test/zdtm/static/Makefile

@@ -377,6 +377,7 @@ TST_DIR		=				\
 		cgroup03			\
 		cgroup04			\
 		cgroup_ifpriomap		\
+		cgroup_ignore			\
 		cgroup_stray			\
 		cgroup_yard			\
 		unlink_fstat04			\

test/zdtm/static/cgroup_ignore.c

@@ -0,0 +1,161 @@
+#include <unistd.h>
+#include <stdio.h>
+#include <string.h>
+#include <fcntl.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/mount.h>
+#include "zdtmtst.h"
+
+const char *test_doc = "Check that cgroups are correctly ignored";
+const char *test_author = "Adrian Reber <areber@redhat.com>";
+
+char *dirname;
+TEST_OPTION(dirname, string, "cgroup directory name", 1);
+static const char *cgname = "zdtmtst";
+
+static size_t read_all(int fd, char *buf, size_t size)
+{
+	ssize_t r = 0, ret;
+
+	while (r < size) {
+		ret = read(fd, buf + r, size - r);
+		if (ret < 0) {
+			pr_perror("Read failed");
+			return -1;
+		} else if (ret == 0) {
+			return 0;
+		}
+		r += ret;
+	}
+
+	return 0;
+}
+
+int main(int argc, char **argv)
+{
+	cleanup_free char *cgroup_procs = NULL;
+	cleanup_close int cgroup_procs_fd = -1;
+	cleanup_free char *destination = NULL;
+	cleanup_free char *buffer_old = NULL;
+	cleanup_free char *buffer_new = NULL;
+	cleanup_close int fd = -1;
+	int ret = 1;
+
+	test_init(argc, argv);
+
+	buffer_old = malloc(PAGE_SIZE);
+	if (!buffer_old) {
+		pr_err("Could not allocate memory\n");
+		return 1;
+	}
+	memset(buffer_old, 0, PAGE_SIZE);
+	buffer_new = malloc(PAGE_SIZE);
+	if (!buffer_new) {
+		pr_err("Could not allocate memory\n");
+		return 1;
+	}
+	memset(buffer_new, 0, PAGE_SIZE);
+
+	// Read /proc/self/cgroup to later compare against it
+	fd = open("/proc/self/cgroup", O_RDONLY);
+	if (fd < 0) {
+		pr_err("Could not open /proc/self/cgroup\n");
+		return 1;
+	}
+
+	if (read_all(fd, buffer_old, PAGE_SIZE)) {
+		pr_err("Could not read data from /proc/self/cgroup\n");
+		return 1;
+	}
+
+	// Create the cgroup root directory
+	if (mkdir(dirname, 0700) < 0 && errno != EEXIST) {
+		pr_err("Cannot make directory %s\n", dirname);
+		return 1;
+	}
+
+	// Mount cgroup2, skip if cgroup2 is not available
+	if (mount("none", dirname, "cgroup2", 0, 0)) {
+		if (errno == ENODEV) {
+			skip("Test relies on cgroup2 semantics which this system does not support. Skipping");
+			test_daemon();
+			test_waitsig();
+			pass();
+			return 0;
+		} else {
+			pr_err("Could not mount cgroup2 at %s\n", dirname);
+		}
+		return 1;
+	}
+
+	// Create the cgroup cgname (if it does not already exist)
+	if (asprintf(&destination, "%s/%s", dirname, cgname) == -1) {
+		pr_err("Could not allocate memory\n");
+		goto err;
+	}
+	if (mkdir(destination, 0700) < 0 && errno != EEXIST) {
+		pr_err("Failed to create temporary cgroup directory %s\n", destination);
+		goto err;
+	}
+
+	// Move this process to the newly created cgroup
+	if (asprintf(&cgroup_procs, "%s/cgroup.procs", destination) == -1) {
+		pr_err("Could not allocate memory\n");
+		goto err;
+	}
+	cgroup_procs_fd = open(cgroup_procs, O_RDWR);
+	if (cgroup_procs_fd < 0) {
+		pr_err("Could not open %s\n", cgroup_procs);
+		goto err;
+	}
+	if (write(cgroup_procs_fd, "0", 1) != 1) {
+		pr_err("Writing to %s failed\n", cgroup_procs);
+		goto err;
+	}
+
+	// Read /proc/self/cgroup (should have changed)
+	lseek(fd, 0, SEEK_SET);
+	if (read_all(fd, buffer_new, PAGE_SIZE)) {
+		pr_err("Could not read data from /proc/self/cgroup\n");
+		goto err;
+	}
+
+	// Test if /proc/self/cgroup has changed
+	if (!memcmp(buffer_new, buffer_old, PAGE_SIZE)) {
+		fail("/proc/self/cgroup should differ after move to another cgroup");
+		pr_err("original /proc/self/cgroup content %s\n", buffer_old);
+		pr_err("new /proc/self/cgroup content %s\n", buffer_new);
+		goto err;
+	}
+
+	test_daemon();
+	test_waitsig();
+
+	// Read /proc/self/cgroup. It should not be the same as after
+	// moving this process to another cgroup because of restore
+	// with '--manage-cgroups=ignore'. The process should be
+	// now in cgroup of the current session.
+	lseek(fd, 0, SEEK_SET);
+	memset(buffer_old, 0, PAGE_SIZE);
+	if (read_all(fd, buffer_old, PAGE_SIZE)) {
+		pr_err("Could not read data from /proc/self/cgroup\n");
+		goto err;
+	}
+
+	// Test if /proc/self/cgroup has changed again
+	if (!memcmp(buffer_new, buffer_old, PAGE_SIZE)) {
+		fail("/proc/self/cgroup should differ after restore");
+		pr_err("original /proc/self/cgroup content %s\n", buffer_new);
+		pr_err("new /proc/self/cgroup content %s\n", buffer_old);
+		goto err;
+	}
+
+	ret = 0;
+	pass();
+err:
+	rmdir(destination);
+	umount(dirname);
+
+	return ret;
+}

test/zdtm/static/cgroup_ignore.checkskip

@@ -0,0 +1,13 @@
+#!/bin/bash
+
+dest=$(mktemp -d cg_ignore.XXXXXX)
+
+trap 'rmdir "$dest"' EXIT
+
+if ! mount -t cgroup2 none "$dest"; then
+	exit 1
+fi
+
+umount "$dest"
+
+exit 0

test/zdtm/static/cgroup_ignore.desc

@@ -0,0 +1 @@
+{'flavor': 'h', 'flags': 'suid excl', 'ropts': '--manage-cgroups=ignore'}