From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from sog-mx-3.v43.ch3.sourceforge.com ([172.29.43.193] helo=mx.sourceforge.net) by sfs-ml-4.v29.ch3.sourceforge.com with esmtp (Exim 4.76) (envelope-from ) id 1UfYOO-0001np-NG for ltp-list@lists.sourceforge.net; Thu, 23 May 2013 16:29:28 +0000 Received: from aserp1040.oracle.com ([141.146.126.69]) by sog-mx-3.v43.ch3.sourceforge.com with esmtps (TLSv1:AES256-SHA:256) (Exim 4.76) id 1UfYOM-00020m-K9 for ltp-list@lists.sourceforge.net; Thu, 23 May 2013 16:29:28 +0000 Received: from ucsinet21.oracle.com (ucsinet21.oracle.com [156.151.31.93]) by aserp1040.oracle.com (Sentrion-MTA-4.3.1/Sentrion-MTA-4.3.1) with ESMTP id r4NGTJlf001551 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=OK) for ; Thu, 23 May 2013 16:29:20 GMT Received: from userz7022.oracle.com (userz7022.oracle.com [156.151.31.86]) by ucsinet21.oracle.com (8.14.4+Sun/8.14.4) with ESMTP id r4NGTJJF025050 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=FAIL) for ; Thu, 23 May 2013 16:29:19 GMT Received: from abhmt103.oracle.com (abhmt103.oracle.com [141.146.116.55]) by userz7022.oracle.com (8.14.4+Sun/8.14.4) with ESMTP id r4NGTIkv029640 for ; Thu, 23 May 2013 16:29:18 GMT From: Alexey Kodanev Date: Thu, 23 May 2013 20:28:46 +0400 Message-Id: <1369326526-20449-1-git-send-email-alexey.kodanev@oracle.com> Subject: [LTP] [PATCH] New core test of cgroups and extended attributes List-Id: Linux Test Project General Discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Errors-To: ltp-list-bounces@lists.sourceforge.net To: ltp-list@lists.sourceforge.net Cc: vasily.isaenko@oracle.com, Alexey Kodanev Support of extended attributes in cgroups was added in Linux 3.7. Signed-off-by: Alexey Kodanev --- runtest/controllers | 1 + .../kernel/controllers/cgroup_xattr/.gitignore | 1 + testcases/kernel/controllers/cgroup_xattr/Makefile | 19 + testcases/kernel/controllers/cgroup_xattr/README | 25 ++ .../kernel/controllers/cgroup_xattr/cgroup_xattr.c | 402 ++++++++++++++++++++ 5 files changed, 448 insertions(+), 0 deletions(-) create mode 100644 testcases/kernel/controllers/cgroup_xattr/.gitignore create mode 100644 testcases/kernel/controllers/cgroup_xattr/Makefile create mode 100644 testcases/kernel/controllers/cgroup_xattr/README create mode 100644 testcases/kernel/controllers/cgroup_xattr/cgroup_xattr.c diff --git a/runtest/controllers b/runtest/controllers index b78d828..94fc185 100644 --- a/runtest/controllers +++ b/runtest/controllers @@ -12,3 +12,4 @@ memcg_stress memcg_stress_test.sh memcg_control PAGESIZE=$(mem_process -p);memcg_control_test.sh $PAGESIZE $PAGESIZE $((PAGESIZE * 2)) cgroup_fj run_cgroup_test_fj.sh controllers test_controllers.sh +cgroup_xattr cgroup_xattr diff --git a/testcases/kernel/controllers/cgroup_xattr/.gitignore b/testcases/kernel/controllers/cgroup_xattr/.gitignore new file mode 100644 index 0000000..3664cc9 --- /dev/null +++ b/testcases/kernel/controllers/cgroup_xattr/.gitignore @@ -0,0 +1 @@ +/cgroup_xattr diff --git a/testcases/kernel/controllers/cgroup_xattr/Makefile b/testcases/kernel/controllers/cgroup_xattr/Makefile new file mode 100644 index 0000000..cd49588 --- /dev/null +++ b/testcases/kernel/controllers/cgroup_xattr/Makefile @@ -0,0 +1,19 @@ +# Copyright (c) 2013 Oracle and/or its affiliates. All Rights Reserved. +# +# This program is free software; you can redistribute it and/or +# modify it under the terms of the GNU General Public License 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. See the +# GNU General Public License for more details. +# +# You should have received a copy of the GNU General Public License +# along with this program; if not, write the Free Software Foundation, +# Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + +top_srcdir ?= ../../../.. + +include $(top_srcdir)/include/mk/testcases.mk +include $(top_srcdir)/include/mk/generic_leaf_target.mk diff --git a/testcases/kernel/controllers/cgroup_xattr/README b/testcases/kernel/controllers/cgroup_xattr/README new file mode 100644 index 0000000..9cc5176 --- /dev/null +++ b/testcases/kernel/controllers/cgroup_xattr/README @@ -0,0 +1,25 @@ +TEST SUITE: + +The directory cgroup_xattr contains the tests related to extended +attributes in cgroup filesystem. + +WARNING: + +This test can cause a kernel panic due to a bug in kernels prior to 3.8. +It was fixed by kernel upstream commit +712317ad97f41e738e1a19aa0a6392a78a84094e: + +"We should store file xattrs in struct cfent instead of struct cftype, +because cftype is a type while cfent is object instance of cftype." + +TESTS AIM: + +The aim of the tests is to check the extended attributes in cgroup +filesystem. This feature was added in Linux 3.7 to allow attaching runtime +meta information to cgroups and everything they model (services, apps, vms) +and can easily be shared among applications. + +Test creates as many subsystems as possible (cpu, cpuset, ...) in the +cgroup tmp directory and in one more created hierarchy. Then sets extended +attributes to all files in cgroup fs and subsequently reads the file's +extended attributes back, checking values during the process. diff --git a/testcases/kernel/controllers/cgroup_xattr/cgroup_xattr.c b/testcases/kernel/controllers/cgroup_xattr/cgroup_xattr.c new file mode 100644 index 0000000..4ef5e16 --- /dev/null +++ b/testcases/kernel/controllers/cgroup_xattr/cgroup_xattr.c @@ -0,0 +1,402 @@ +/* + * Copyright (c) 2013 Oracle and/or its affiliates. All Rights Reserved. + * + * This program is free software; you can redistribute it and/or + * modify it under the terms of the GNU General Public License 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. See the + * GNU General Public License for more details. + * + * You should have received a copy of the GNU General Public License + * along with this program; if not, write the Free Software Foundation, + * Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA + * + * Author: + * Alexey Kodanev + * + * Test checks following preconditions: + * since Linux kernel 3.7 it is possible to set extended attributes + * to cgroup files. + */ + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "test.h" +#include "usctest.h" +#include "safe_macros.h" + +char *TCID = "cgroup_xattr"; + +static const char cgrp_point[] = "cgroup"; +static const char cgrp_name[] = "cgrp_test"; +static const char subdir_name[] = "test"; + +#define MAX_SUBSYS 64 +#define MAX_OPTIONS_LEN 256 + +static char subsys[MAX_OPTIONS_LEN]; + +struct tst_key { + const char *name; + int good; +}; + +/* only security.* & trusted.* are valid key names */ +static const struct tst_key tkeys[] = { + { .name = "trusted.test", .good = 1, }, + { .name = "security.", .good = 1, }, + { .name = "user.", .good = 0, }, + { .name = "system.", .good = 0, }, +}; + +#define DEFAULT_VALUE_SIZE 8 + +/* struct to store key's value */ +struct tst_val { + char *buf; + size_t size; +}; +static struct tst_val val; + +/* it fills value's buffer */ +static char id; + +/* cleanup flags */ +static int cgrp_mounted; +static int subdir_created; + +/* + * When test breaks, all open dirs should be closed + * otherwise umount won't succeed + */ +#define MAX_OPEN_DIR 32 +static DIR *odir[MAX_OPEN_DIR]; +static int odir_num; + +/* test options */ +static char *narg; +static int nflag; +static int skip_cleanup; +static int verbose; +static const option_t options[] = { + {"n:", &nflag, &narg}, + {"s", &skip_cleanup, NULL}, + {"v", &verbose, NULL}, + {NULL, NULL, NULL} +}; + +static void help(void); +static void setup(int argc, char *argv[]); +static void test_run(void); +static void cleanup(void); + +/* + * get info about mounted subsystems, + * it is possible to mount only already mounted subsystems + * or ones not mounted, other combinations won't succeed + */ +static void make_cgroup_options(void); +static int set_xattrs(const char *file); +static int get_xattrs(const char *file); +/* + * set or get xattr recursively + * + * @path: start directory + * @xattr_operation: can be set_xattrs() or get_xattrs() + */ +static int cgrp_files_walking(const char *path, + int (*xattr_operation)(const char *)); + +int main(int argc, char *argv[]) +{ + setup(argc, argv); + + test_run(); + + cleanup(); + + tst_exit(); +} + +static void help(void) +{ + printf(" -n x Write x bytes to xattr value, default is %d\n", + DEFAULT_VALUE_SIZE); + printf(" -s Skip cleanup\n"); + printf(" -v Verbose\n"); +} + +void setup(int argc, char *argv[]) +{ + char *msg; + msg = parse_opts(argc, argv, options, help); + if (msg != NULL) + tst_brkm(TBROK, NULL, "OPTION PARSING ERROR - %s", msg); + + tst_require_root(NULL); + + if (access("/proc/cgroups", F_OK) == -1) + tst_brkm(TCONF, NULL, "Kernel doesn't support for cgroups"); + + if (tst_kvercmp(3, 7, 0) < 0) { + tst_brkm(TCONF, NULL, + "Test must be run with kernel 3.7 or newer"); + } + + int value_size = DEFAULT_VALUE_SIZE; + if (nflag) { + if (sscanf(narg, "%i", &value_size) != 1) + tst_brkm(TBROK, NULL, "-n option arg is not a number"); + if (value_size <= 0) + tst_brkm(TBROK, NULL, "-n option arg is less than 1"); + } + + /* initialize test value */ + val.size = value_size; + val.buf = SAFE_MALLOC(NULL, value_size); + + tst_sig(FORK, DEF_HANDLER, cleanup); + + tst_tmpdir(); + + SAFE_MKDIR(cleanup, cgrp_point, 0755); + + make_cgroup_options(); + + if (mount(cgrp_name, cgrp_point, "cgroup", 0, subsys) == -1) + tst_brkm(TBROK, cleanup, "Can't mount: %s", cgrp_point); + cgrp_mounted = 1; + + /* create new hierarchy */ + SAFE_CHDIR(cleanup, cgrp_point); + SAFE_MKDIR(cleanup, subdir_name, 0755); + subdir_created = 1; +} + +static void test_run(void) +{ + /* set xattr to each file in cgroup fs */ + int set_res = cgrp_files_walking(".", set_xattrs); + /* reset value */ + id = 0; + /* get & check xattr */ + int get_res = cgrp_files_walking(".", get_xattrs); + + /* final results */ + tst_resm(TINFO, "All test-cases have been completed, summary:"); + tst_resm(TINFO, "Set tests result: %s", (set_res) ? "FAIL" : "PASS"); + tst_resm(TINFO, "Get tests result: %s", (get_res) ? "FAIL" : "PASS"); +} + +static void cleanup(void) +{ + if (val.buf != NULL) + free(val.buf); + + if (skip_cleanup) + return; + + /* + * Kernels 3.7 can crash while unmounting cgroups with xattr, + * call tst_flush() to make sure all buffered data written + * before it happens + */ + tst_flush(); + + int i; + for (i = 0; i < odir_num; ++i) { + if (closedir(odir[i]) == -1) + tst_brkm(TBROK, NULL, "Failed to close dir\n"); + } + + char *cwd = get_tst_tmpdir(); + SAFE_CHDIR(NULL, cwd); + free(cwd); + + if (subdir_created) { + SAFE_CHDIR(NULL, cgrp_point); + if (rmdir(subdir_name) == -1) + tst_brkm(TBROK | TERRNO, NULL, "Can't remove dir"); + SAFE_CHDIR(NULL, ".."); + } + + if (cgrp_mounted) { + if (umount(cgrp_point) == -1) { + tst_brkm(TBROK | TERRNO, NULL, + "Can't unmount: %s", cgrp_point); + } + } + + tst_rmdir(); + TEST_CLEANUP; +} + +void make_cgroup_options(void) +{ + /* detect sybsystems */ + FILE *fd = fopen("/proc/cgroups", "r"); + if (fd == NULL) + tst_brkm(TBROK, cleanup, "Failed to read /proc/cgroups"); + + int subsys_ynum = 0; + /* store not mounted subsys here*/ + char subsys_n[MAX_OPTIONS_LEN]; + int subsys_nnum = 0; + + strcpy(subsys, "xattr"); + strcpy(subsys_n, "xattr"); + + char str[MAX_SUBSYS]; + + int first = 1; + + while ((fgets(str, MAX_SUBSYS, fd)) != NULL) { + /* skip first line */ + if (first) { + first = 0; + continue; + } + int num = 0; + char name[MAX_SUBSYS]; + sscanf(str, "%s\t%d", name, &num); + + strcat((num > 0) ? subsys : subsys_n, ","); + strcat((num > 0) ? subsys : subsys_n, name); + + if (num > 0) + ++subsys_ynum; + else + ++subsys_nnum; + } + fclose(fd); + + /* get as many subsystems as possible */ + if (subsys_ynum < subsys_nnum) + strcpy(subsys, subsys_n); + + tst_resm(TINFO, "mount options: %s", subsys); +} + +static int set_xattrs(const char *file) +{ + int i, err, fail, res; + res = 0; + for (i = 0; i < ARRAY_SIZE(tkeys); ++i) { + err = setxattr(file, tkeys[i].name, + (const void *)val.buf, val.size, 0) == -1; + + fail = err && tkeys[i].good; + res |= fail; + + tst_resm((fail) ? TFAIL : TPASS, + "Expect: %s set xattr key '%s' to file '%s'", + (tkeys[i].good) ? "can" : "can't", + tkeys[i].name, file); + + if (verbose && tkeys[i].good) + tst_resm_hexd(TINFO, val.buf, val.size, "value:"); + } + return res; +} + +static int get_xattrs(const char *file) +{ + int i, fail, res; + res = 0; + for (i = 0; i < ARRAY_SIZE(tkeys); ++i) { + /* get value size */ + ssize_t size = getxattr(file, tkeys[i].name, NULL, 0); + fail = (size == -1 && tkeys[i].good); + res |= fail; + + tst_resm((fail) ? TFAIL : TPASS, + "Expect: %s read xattr %s of file '%s'", + (size == -1) ? "can't" : "can", + tkeys[i].name, file); + + if (fail || size == -1) + continue; + + /* get xattr value */ + char xval[size]; + if (getxattr(file, tkeys[i].name, xval, size) == -1) { + tst_brkm(TBROK, cleanup, + "Can't get buffer of key %s", + tkeys[i].name); + } + fail = val.size != size || + strncmp(val.buf, xval, val.size) != 0; + res |= fail; + + tst_resm((fail) ? TFAIL : TPASS, "Expect: values equal"); + + if (verbose && fail) { + tst_resm_hexd(TINFO, xval, size, + "Read xattr value:"); + tst_resm_hexd(TINFO, val.buf, val.size, + "Expect xattr value:"); + } + } + return res; +} + +static int cgrp_files_walking(const char *path, + int (*xattr_operation)(const char *)) +{ + int res = 0; + struct dirent *entry; + DIR *dir = opendir(path); + + odir[odir_num] = dir; + if (++odir_num >= MAX_OPEN_DIR) { + tst_brkm(TBROK, cleanup, + "Unexpected num of open dirs, max: %d", MAX_OPEN_DIR); + } + + SAFE_CHDIR(cleanup, path); + + tst_resm(TINFO, "In dir %s", path); + + errno = 0; + while ((entry = readdir(dir)) != NULL) { + const char *file = entry->d_name; + /* skip current and up directories */ + if (!strcmp(file, "..") || !strcmp(file, ".")) + continue; + struct stat stat_buf; + TEST(lstat(file, &stat_buf)); + if (TEST_RETURN != -1 && S_ISDIR(stat_buf.st_mode)) { + /* proceed to subdir */ + res |= cgrp_files_walking(file, xattr_operation); + tst_resm(TINFO, "In dir %s", path); + } + memset(val.buf, id++, val.size); + res |= xattr_operation(file); + errno = 0; + } + if (errno && !entry) { + tst_brkm(TWARN | TERRNO, cleanup, + "Error while reading dir '%s'", path); + } + if (closedir(dir) == -1) + tst_brkm(TWARN, cleanup, "Failed to close dir '%s'", path); + else + odir[--odir_num] = NULL; + + if (strcmp(path, ".")) + SAFE_CHDIR(cleanup, ".."); + return res; +} -- 1.7.1 ------------------------------------------------------------------------------ Try New Relic Now & We'll Send You this Cool Shirt New Relic is the only SaaS-based application performance monitoring service that delivers powerful full stack analytics. Optimize and monitor your browser, app, & servers with just a few lines of code. Try New Relic and get this awesome Nerd Life shirt! http://p.sf.net/sfu/newrelic_d2d_may _______________________________________________ Ltp-list mailing list Ltp-list@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ltp-list