openkylin
/
libvirt
mirror of https://gitee.com/openkylin/libvirt.git
9
0
Fork 0
Code Issues Projects Releases Wiki Activity
libvirt/tests/vircgrouptest.c

1139 lines
39 KiB
C
Raw Normal View History

Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
/*
tests: fix build on old 32-bit platforms gcc 4.1.2 (hello RHEL 5) on 32-bit platforms complains: vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:627: warning: integer constant is too large for 'long' type vircgrouptest.c:628: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:634: warning: integer constant is too large for 'long' type vircgrouptest.c:635: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:636: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:644: warning: integer constant is too large for 'long' type * tests/vircgrouptest.c (testCgroupGetPercpuStats): Use ULL suffix. Signed-off-by: Eric Blake <eblake@redhat.com>
2015-04-16 02:50:23 +08:00
* Copyright (C) 2013-2015 Red Hat, Inc.
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation; either
* version 2.1 of the License, or (at your option) any later version.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library. If not, see
* <http://www.gnu.org/licenses/>.
*/
#include <config.h>
#include "testutils.h"
#ifdef __linux__
Enforce a standard header file guard symbol name Require that all headers are guarded by a symbol named LIBVIRT_$FILENAME where $FILENAME is the uppercased filename, with all characters outside a-z changed into '_'. Note we do not use a leading __ because that is technically a namespace reserved for the toolchain. Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2018-12-13 22:53:50 +08:00
# define LIBVIRT_VIRCGROUPPRIV_H_ALLOW
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
# include "vircgrouppriv.h"
virutil: Move string related functions to virstring.c The source code base needs to be adapted as well. Some files include virutil.h just for the string related functions (here, the include is substituted to match the new file), some include virutil.h without any need (here, the include is removed), and some require both.
2013-04-03 18:36:23 +08:00
# include "virstring.h"
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
# include "virerror.h"
# include "virlog.h"
# include "virfile.h"
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
# include "virbuffer.h"
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
# include "testutilslxc.h"
nodeinfo: move host CPU APIs out into virhostcpu.c file Move all APIs with a virHostCPU name prefix out into new util/virhostcpu.h & util/virhostcpu.c files Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2016-04-14 01:53:02 +08:00
# include "virhostcpu.h"
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
# define VIR_FROM_THIS VIR_FROM_NONE
Add virLogSource variables to all source files Any source file which calls the logging APIs now needs to have a VIR_LOG_INIT("source.name") declaration at the start of the file. This provides a static variable of the virLogSource type. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2014-02-28 20:16:17 +08:00
VIR_LOG_INIT("tests.cgrouptest");
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
static int validateCgroup(virCgroupPtr cgroup,
const char *expectPath,
const char **expectMountPoint,
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
const char **expectLinkPoint,
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
const char **expectPlacement,
const char *expectUnifiedMountPoint,
const char *expectUnifiedPlacement,
unsigned int expectUnifiedControllers)
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
{
Convert 'int i' to 'size_t i' in tests/ files Convert the type of loop iterators named 'i', 'j', k', 'ii', 'jj', 'kk', to be 'size_t' instead of 'int' or 'unsigned int', also santizing 'ii', 'jj', 'kk' to use the normal 'i', 'j', 'k' naming Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-08 22:09:33 +08:00
size_t i;
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
if (STRNEQ(cgroup->path, expectPath)) {
fprintf(stderr, "Wrong path '%s', expected '%s'\n",
cgroup->path, expectPath);
return -1;
}
tests/: Remove the whitespace before ";"
2013-05-21 15:53:48 +08:00
for (i = 0; i < VIR_CGROUP_CONTROLLER_LAST; i++) {
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
if (STRNEQ_NULLABLE(expectMountPoint[i],
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].mountPoint)) {
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
fprintf(stderr, "Wrong mount '%s', expected '%s' for '%s'\n",
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].mountPoint,
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
expectMountPoint[i],
virCgroupControllerTypeToString(i));
return -1;
}
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
if (STRNEQ_NULLABLE(expectLinkPoint[i],
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].linkPoint)) {
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
fprintf(stderr, "Wrong link '%s', expected '%s' for '%s'\n",
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].linkPoint,
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
expectLinkPoint[i],
virCgroupControllerTypeToString(i));
return -1;
}
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
if (STRNEQ_NULLABLE(expectPlacement[i],
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].placement)) {
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
fprintf(stderr, "Wrong placement '%s', expected '%s' for '%s'\n",
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
cgroup->legacy[i].placement,
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
expectPlacement[i],
virCgroupControllerTypeToString(i));
return -1;
}
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
if (STRNEQ_NULLABLE(expectUnifiedMountPoint,
cgroup->unified.mountPoint)) {
fprintf(stderr, "Wrong mount '%s', expected '%s' for 'unified'\n",
cgroup->unified.mountPoint,
expectUnifiedMountPoint);
return -1;
}
if (STRNEQ_NULLABLE(expectUnifiedPlacement,
cgroup->unified.placement)) {
fprintf(stderr, "Wrong placement '%s', expected '%s' for 'unified'\n",
cgroup->unified.placement,
expectUnifiedPlacement);
return -1;
}
if (expectUnifiedControllers != cgroup->unified.controllers) {
for (i = 0; i < VIR_CGROUP_CONTROLLER_LAST; i++) {
int type = 1 << i;
if ((expectUnifiedControllers & type) != (cgroup->unified.controllers & type)) {
const char *typeStr = virCgroupControllerTypeToString(i);
if (expectUnifiedControllers & type) {
fprintf(stderr, "expected controller '%s' for 'unified', "
"but it's missing\n", typeStr);
} else {
fprintf(stderr, "existing controller '%s' for 'unified', "
"but it's not expected\n", typeStr);
}
}
}
return -1;
}
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
return 0;
}
const char *mountsSmall[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/not/really/sys/fs/cgroup/cpu,cpuacct",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/not/really/sys/fs/cgroup/cpu,cpuacct",
[VIR_CGROUP_CONTROLLER_CPUSET] = NULL,
[VIR_CGROUP_CONTROLLER_MEMORY] = "/not/really/sys/fs/cgroup/memory",
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = NULL,
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL,
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
};
const char *mountsFull[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/not/really/sys/fs/cgroup/cpu,cpuacct",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/not/really/sys/fs/cgroup/cpu,cpuacct",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/not/really/sys/fs/cgroup/cpuset",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/not/really/sys/fs/cgroup/memory",
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = "/not/really/sys/fs/cgroup/freezer",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/not/really/sys/fs/cgroup/blkio",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/not/really/sys/fs/cgroup/systemd",
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
};
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
const char *mountsAllInOne[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_DEVICES] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = "/not/really/sys/fs/cgroup",
[VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL,
};
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
const char *links[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/not/really/sys/fs/cgroup/cpu",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/not/really/sys/fs/cgroup/cpuacct",
[VIR_CGROUP_CONTROLLER_CPUSET] = NULL,
[VIR_CGROUP_CONTROLLER_MEMORY] = NULL,
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = NULL,
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL,
};
const char *linksAllInOne[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = NULL,
[VIR_CGROUP_CONTROLLER_CPUACCT] = NULL,
[VIR_CGROUP_CONTROLLER_CPUSET] = NULL,
[VIR_CGROUP_CONTROLLER_MEMORY] = NULL,
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = NULL,
[VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL,
Track symlinks for co-mounted cgroup controllers If a cgroup controller is co-mounted with another, eg /sys/fs/cgroup/cpu,cpuacct Then it is a requirement that there exist symlinks at /sys/fs/cgroup/cpu /sys/fs/cgroup/cpuacct pointing to the real mount point. Add support to virCgroupPtr to detect and track these symlinks Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-05 18:28:04 +08:00
};
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
struct _detectMountsData {
const char *file;
bool fail;
};
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
static int
testCgroupDetectMounts(const void *args)
{
int result = -1;
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
const struct _detectMountsData *data = args;
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
char *parsed = NULL;
const char *actual;
virCgroupPtr group = NULL;
tests: use g_auto for all virBuffers Signed-off-by: Laine Stump <laine@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
2020-07-03 07:35:41 +08:00
g_auto(virBuffer) buf = VIR_BUFFER_INITIALIZER;
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
size_t i;
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_setenv("VIR_CGROUP_MOCK_FILENAME", data->file, TRUE);
vircgrouptest: call virCgroupDetectMounts directly Because we can set which files to return for cgroup tests there is no need to have special function tailored to run tests. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 18:38:54 +08:00
tests: Use g_strdup_printf() instead of virAsprintf() Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Daniel Henrique Barboza <danielhb413@gmail.com>
2019-10-22 21:26:14 +08:00
parsed = g_strdup_printf("%s/vircgroupdata/%s.parsed", abs_srcdir, data->file);
vircgrouptest: call virCgroupDetectMounts directly Because we can set which files to return for cgroup tests there is no need to have special function tailored to run tests. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 18:38:54 +08:00
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
if (virCgroupNewSelf(&group) < 0) {
if (data->fail)
result = 0;
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
goto cleanup;
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
}
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
if (data->fail)
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
goto cleanup;
for (i = 0; i < VIR_CGROUP_CONTROLLER_LAST; i++) {
virBufferAsprintf(&buf, "%-12s %s\n",
virCgroupControllerTypeToString(i),
vircgroup: rename controllers to legacy With the introduction of cgroup v2 there are new names used with cgroups based on which version is used: - legacy: cgroup v1 - unified: cgroup v2 - hybrid: cgroup v1 and cgroup v2 Let's use 'legacy' instead of 'cgroupv1' or 'controllers' in our code. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-25 07:17:02 +08:00
NULLSTR(group->legacy[i].mountPoint));
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
}
vircgrouptest: prepare testCgroupDetectMounts for cgroup v2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:28:04 +08:00
virBufferAsprintf(&buf, "%-12s %s\n",
"unified", NULLSTR(group->unified.mountPoint));
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
actual = virBufferCurrentContent(&buf);
tests: Rename virtTestCompareToFile to virTestCompareToFile. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:53 +08:00
if (virTestCompareToFile(actual, parsed) < 0)
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
goto cleanup;
result = 0;
cleanup:
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_unsetenv("VIR_CGROUP_MOCK_FILENAME");
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
VIR_FREE(parsed);
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&group);
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
return result;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForSelf(const void *args G_GNUC_UNUSED)
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
const char *placement[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/system",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/system",
Refactor cgroups internal data structures Currently the virCgroupPtr struct contains 3 pieces of information - path - path of the cgroup, relative to current process' cgroup placement - placement - current process' placement in each controller - mounts - mount point of each controller When reading/writing cgroup settings, the path & placement strings are combined to form the file path. This approach only works if we assume all cgroups will be relative to the current process' cgroup placement. To allow support for managing cgroups at any place in the heirarchy a change is needed. The 'placement' data should reflect the absolute path to the cgroup, and the 'path' value should no longer be used to form the paths to the cgroup attribute files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-22 19:11:34 +08:00
[VIR_CGROUP_CONTROLLER_CPUSET] = "/",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/",
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
Refactor cgroups internal data structures Currently the virCgroupPtr struct contains 3 pieces of information - path - path of the cgroup, relative to current process' cgroup placement - placement - current process' placement in each controller - mounts - mount point of each controller When reading/writing cgroup settings, the path & placement strings are combined to form the file path. This approach only works if we assume all cgroups will be relative to the current process' cgroup placement. To allow support for managing cgroups at any place in the heirarchy a change is needed. The 'placement' data should reflect the absolute path to the cgroup, and the 'path' value should no longer be used to form the paths to the cgroup attribute files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-22 19:11:34 +08:00
[VIR_CGROUP_CONTROLLER_FREEZER] = "/",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
};
if (virCgroupNewSelf(&cgroup) < 0) {
fprintf(stderr, "Cannot create cgroup for self\n");
goto cleanup;
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(cgroup, "", mountsFull, links, placement, NULL, NULL, 0);
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
return ret;
}
Remove backslash alignment attempts Right-aligning backslashes when defining macros or using complex commands in Makefiles looks cute, but as soon as any changes is required to the code you end up with either distractingly broken alignment or unnecessarily big diffs where most of the changes are just pushing all backslashes a few characters to one side. Generated using $ git grep -El '[[:blank:]][[:blank:]]\\$' | \ grep -E '*\.([chx]|am|mk)$$' | \ while read f; do \ sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \ done Signed-off-by: Andrea Bolognani <abologna@redhat.com>
2017-11-03 20:09:47 +08:00
# define ENSURE_ERRNO(en) \
do { \
if (!virLastErrorIsSystemErrno(en)) { \
virErrorPtr err = virGetLastError(); \
fprintf(stderr, "Did not get " #en " error code: %d:%d\n", \
err ? err->code : 0, err ? err->int1 : 0); \
goto cleanup; \
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
} } while (0)
/* Asking for impossible combination since CPU is co-mounted */
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForPartition(const void *args G_GNUC_UNUSED)
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
int rv;
const char *placementSmall[VIR_CGROUP_CONTROLLER_LAST] = {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_CPU] = "/virtualmachines.partition",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/virtualmachines.partition",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
[VIR_CGROUP_CONTROLLER_CPUSET] = NULL,
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_MEMORY] = "/virtualmachines.partition",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = NULL,
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = NULL,
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
};
const char *placementFull[VIR_CGROUP_CONTROLLER_LAST] = {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_CPU] = "/virtualmachines.partition",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/virtualmachines.partition",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/virtualmachines.partition",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/virtualmachines.partition",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_FREEZER] = "/virtualmachines.partition",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/virtualmachines.partition",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
};
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", false, -1, &cgroup)) != -1) {
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
fprintf(stderr, "Unexpected found /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENOENT);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
/* Asking for impossible combination since CPU is co-mounted */
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_CPU),
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
&cgroup)) != -1) {
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
fprintf(stderr, "Should not have created /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(EINVAL);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
/* Asking for impossible combination since devices is not mounted */
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_DEVICES),
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
&cgroup)) != -1) {
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
fprintf(stderr, "Should not have created /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENXIO);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
/* Asking for small combination since devices is not mounted */
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_CPU) |
(1 << VIR_CGROUP_CONTROLLER_CPUACCT) |
(1 << VIR_CGROUP_CONTROLLER_MEMORY),
&cgroup)) != 0) {
fprintf(stderr, "Cannot create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(cgroup, "/virtualmachines.partition", mountsSmall, links, placementSmall, NULL, NULL, 0);
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Cannot create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(cgroup, "/virtualmachines.partition", mountsFull, links, placementFull, NULL, NULL, 0);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
return ret;
}
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForPartitionNested(const void *args G_GNUC_UNUSED)
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
int rv;
const char *placementFull[VIR_CGROUP_CONTROLLER_LAST] = {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_CPU] = "/deployment.partition/production.partition",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/deployment.partition/production.partition",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/deployment.partition/production.partition",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/deployment.partition/production.partition",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_FREEZER] = "/deployment.partition/production.partition",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/deployment.partition/production.partition",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
};
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
if ((rv = virCgroupNewPartition("/deployment/production", false, -1, &cgroup)) != -1) {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
fprintf(stderr, "Unexpected found /deployment/production cgroup: %d\n", -rv);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENOENT);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
/* Should not work, since we require /deployment to be pre-created */
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
if ((rv = virCgroupNewPartition("/deployment/production", true, -1, &cgroup)) != -1) {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
fprintf(stderr, "Unexpected created /deployment/production cgroup: %d\n", -rv);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENOENT);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
if ((rv = virCgroupNewPartition("/deployment", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Failed to create /deployment cgroup: %d\n", -rv);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
goto cleanup;
}
/* Should now work */
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
if ((rv = virCgroupNewPartition("/deployment/production", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Failed to create /deployment/production cgroup: %d\n", -rv);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
goto cleanup;
}
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
ret = validateCgroup(cgroup, "/deployment.partition/production.partition",
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
mountsFull, links, placementFull, NULL, NULL, 0);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForPartitionNestedDeep(const void *args G_GNUC_UNUSED)
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
int rv;
const char *placementFull[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/user/berrange.user/production.partition",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/user/berrange.user/production.partition",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/user/berrange.user/production.partition",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/user/berrange.user/production.partition",
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = "/user/berrange.user/production.partition",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/user/berrange.user/production.partition",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
};
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
if ((rv = virCgroupNewPartition("/user/berrange.user/production", false, -1, &cgroup)) != -1) {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
fprintf(stderr, "Unexpected found /user/berrange.user/production cgroup: %d\n", -rv);
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENOENT);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
/* Should not work, since we require /user/berrange.user to be pre-created */
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
if ((rv = virCgroupNewPartition("/user/berrange.user/production", true, -1, &cgroup)) != -1) {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
fprintf(stderr, "Unexpected created /user/berrange.user/production cgroup: %d\n", -rv);
goto cleanup;
}
Report full errors from virCgroupNew* Instead of returning raw errno values, report full libvirt errors in virCgroupNew* functions. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-04 23:49:24 +08:00
ENSURE_ERRNO(ENOENT);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
if ((rv = virCgroupNewPartition("/user", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Failed to create /user/berrange.user cgroup: %d\n", -rv);
goto cleanup;
}
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
if ((rv = virCgroupNewPartition("/user/berrange.user", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Failed to create /user/berrange.user cgroup: %d\n", -rv);
goto cleanup;
}
/* Should now work */
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
if ((rv = virCgroupNewPartition("/user/berrange.user/production", true, -1, &cgroup)) != 0) {
fprintf(stderr, "Failed to create /user/berrange.user/production cgroup: %d\n", -rv);
goto cleanup;
}
ret = validateCgroup(cgroup, "/user/berrange.user/production.partition",
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
mountsFull, links, placementFull, NULL, NULL, 0);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForPartitionDomain(const void *args G_GNUC_UNUSED)
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
{
virCgroupPtr partitioncgroup = NULL;
virCgroupPtr domaincgroup = NULL;
int ret = -1;
int rv;
const char *placement[VIR_CGROUP_CONTROLLER_LAST] = {
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_CPU] = "/production.partition/foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/production.partition/foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/production.partition/foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/production.partition/foo.libvirt-lxc",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
[VIR_CGROUP_CONTROLLER_FREEZER] = "/production.partition/foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/production.partition/foo.libvirt-lxc",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
};
if ((rv = virCgroupNewPartition("/production", true, -1, &partitioncgroup)) != 0) {
fprintf(stderr, "Failed to create /production cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupNewDomainPartition(partitioncgroup, "lxc", "foo", true, &domaincgroup)) != 0) {
fprintf(stderr, "Cannot create LXC cgroup: %d\n", -rv);
goto cleanup;
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(domaincgroup, "/production.partition/foo.libvirt-lxc", mountsFull, links, placement, NULL, NULL, 0);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&partitioncgroup);
virCgroupFree(&domaincgroup);
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForPartitionDomainEscaped(const void *args G_GNUC_UNUSED)
Do proper escaping of cgroup resource partitions If a user cgroup name begins with "cgroup.", "_" or with any of the controllers from /proc/cgroups followed by a dot, then they need to be prefixed with a single underscore. eg if there is an object "cpu.service", then this would end up as "_cpu.service" in the cgroup filesystem tree, however, "waldo.service" would stay "waldo.service", at least as long as nobody comes up with a cgroup controller called "waldo". Since we require a '.XXXX' suffix on all partitions, there is no scope for clashing with the kernel 'tasks' and 'release_agent' files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:50:24 +08:00
{
virCgroupPtr partitioncgroup1 = NULL;
virCgroupPtr partitioncgroup2 = NULL;
virCgroupPtr partitioncgroup3 = NULL;
virCgroupPtr domaincgroup = NULL;
int ret = -1;
int rv;
const char *placement[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_DEVICES] = NULL,
[VIR_CGROUP_CONTROLLER_FREEZER] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
[VIR_CGROUP_CONTROLLER_BLKIO] = "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc",
Add support for systemd cgroup mount Systemd uses a named cgroup mount for tracking processes. Add it as another type of controller, albeit one which we have to special case in a number of places. In particular we must never create/delete directories there, nor add tasks. Essentially the systemd mount is to be considered read-only for libvirt. With this change both the virCgroupDetectPlacement and virCgroupCopyPlacement methods must be invoked. The copy placement method will copy setup for resource controllers only. The detect placement method will probe for any named controllers, or resource controllers not already setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-07-26 02:13:44 +08:00
[VIR_CGROUP_CONTROLLER_SYSTEMD] = "/user/berrange/123",
Do proper escaping of cgroup resource partitions If a user cgroup name begins with "cgroup.", "_" or with any of the controllers from /proc/cgroups followed by a dot, then they need to be prefixed with a single underscore. eg if there is an object "cpu.service", then this would end up as "_cpu.service" in the cgroup filesystem tree, however, "waldo.service" would stay "waldo.service", at least as long as nobody comes up with a cgroup controller called "waldo". Since we require a '.XXXX' suffix on all partitions, there is no scope for clashing with the kernel 'tasks' and 'release_agent' files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:50:24 +08:00
};
if ((rv = virCgroupNewPartition("/cgroup.evil", true, -1, &partitioncgroup1)) != 0) {
fprintf(stderr, "Failed to create /cgroup.evil cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupNewPartition("/cgroup.evil/net_cls.evil", true, -1, &partitioncgroup2)) != 0) {
fprintf(stderr, "Failed to create /cgroup.evil/cpu.evil cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupNewPartition("/cgroup.evil/net_cls.evil/_evil.evil", true, -1, &partitioncgroup3)) != 0) {
fprintf(stderr, "Failed to create /cgroup.evil cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupNewDomainPartition(partitioncgroup3, "lxc", "cpu.foo", true, &domaincgroup)) != 0) {
fprintf(stderr, "Cannot create LXC cgroup: %d\n", -rv);
goto cleanup;
}
/* NB we're not expecting 'net_cls.evil' to be escaped,
* since our fake /proc/cgroups pretends this controller
* isn't compiled into the kernel
*/
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(domaincgroup, "/_cgroup.evil/net_cls.evil/__evil.evil/_cpu.foo.libvirt-lxc", mountsFull, links, placement, NULL, NULL, 0);
Do proper escaping of cgroup resource partitions If a user cgroup name begins with "cgroup.", "_" or with any of the controllers from /proc/cgroups followed by a dot, then they need to be prefixed with a single underscore. eg if there is an object "cpu.service", then this would end up as "_cpu.service" in the cgroup filesystem tree, however, "waldo.service" would stay "waldo.service", at least as long as nobody comes up with a cgroup controller called "waldo". Since we require a '.XXXX' suffix on all partitions, there is no scope for clashing with the kernel 'tasks' and 'release_agent' files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:50:24 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&partitioncgroup3);
virCgroupFree(&partitioncgroup2);
virCgroupFree(&partitioncgroup1);
virCgroupFree(&domaincgroup);
Do proper escaping of cgroup resource partitions If a user cgroup name begins with "cgroup.", "_" or with any of the controllers from /proc/cgroups followed by a dot, then they need to be prefixed with a single underscore. eg if there is an object "cpu.service", then this would end up as "_cpu.service" in the cgroup filesystem tree, however, "waldo.service" would stay "waldo.service", at least as long as nobody comes up with a cgroup controller called "waldo". Since we require a '.XXXX' suffix on all partitions, there is no scope for clashing with the kernel 'tasks' and 'release_agent' files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:50:24 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForSelfAllInOne(const void *args G_GNUC_UNUSED)
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
const char *placement[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPU] = "/",
[VIR_CGROUP_CONTROLLER_CPUACCT] = "/",
[VIR_CGROUP_CONTROLLER_CPUSET] = "/",
[VIR_CGROUP_CONTROLLER_MEMORY] = "/",
[VIR_CGROUP_CONTROLLER_DEVICES] = "/",
[VIR_CGROUP_CONTROLLER_FREEZER] = NULL,
[VIR_CGROUP_CONTROLLER_BLKIO] = "/",
};
if (virCgroupNewSelf(&cgroup) < 0) {
fprintf(stderr, "Cannot create cgroup for self\n");
goto cleanup;
}
vircgrouptest: prepare validateCgroup for cgroupv2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:41:44 +08:00
ret = validateCgroup(cgroup, "", mountsAllInOne, linksAllInOne, placement, NULL, NULL, 0);
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForSelfLogind(const void *args G_GNUC_UNUSED)
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
{
virCgroupPtr cgroup = NULL;
if (virCgroupNewSelf(&cgroup) >= 0) {
fprintf(stderr, "Expected to fail, only systemd cgroup available.\n");
virCgroupFree(&cgroup);
return -1;
}
return 0;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForSelfUnified(const void *args G_GNUC_UNUSED)
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
const char *empty[VIR_CGROUP_CONTROLLER_LAST] = { 0 };
unsigned int controllers =
(1 << VIR_CGROUP_CONTROLLER_CPU) |
(1 << VIR_CGROUP_CONTROLLER_CPUACCT) |
(1 << VIR_CGROUP_CONTROLLER_MEMORY) |
vircgroupmock: mock virCgroupV2DevicesAvailable We need to mock virCgroupV2DevicesAvailable() in order to remove any dependency on kernel as BPF devices might not be available. Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
2019-06-24 21:13:37 +08:00
(1 << VIR_CGROUP_CONTROLLER_DEVICES) |
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
(1 << VIR_CGROUP_CONTROLLER_BLKIO);
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
if (virCgroupNewSelf(&cgroup) < 0) {
fprintf(stderr, "Cannot create cgroup for self\n");
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
goto cleanup;
}
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
ret = validateCgroup(cgroup, "", empty, empty, empty,
"/not/really/sys/fs/cgroup", "/", controllers);
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupNewForSelfHybrid(const void *args G_GNUC_UNUSED)
vircgrouptest: add hybrid tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:47 +08:00
{
virCgroupPtr cgroup = NULL;
int ret = -1;
const char *empty[VIR_CGROUP_CONTROLLER_LAST] = { 0 };
const char *mounts[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPUSET] = "/not/really/sys/fs/cgroup/cpuset",
[VIR_CGROUP_CONTROLLER_FREEZER] = "/not/really/sys/fs/cgroup/freezer",
[VIR_CGROUP_CONTROLLER_NET_CLS] = "/not/really/sys/fs/cgroup/net_cls",
[VIR_CGROUP_CONTROLLER_PERF_EVENT] = "/not/really/sys/fs/cgroup/perf_event",
};
const char *placement[VIR_CGROUP_CONTROLLER_LAST] = {
[VIR_CGROUP_CONTROLLER_CPUSET] = "/",
[VIR_CGROUP_CONTROLLER_FREEZER] = "/",
[VIR_CGROUP_CONTROLLER_NET_CLS] = "/",
[VIR_CGROUP_CONTROLLER_PERF_EVENT] = "/",
};
unsigned int controllers =
(1 << VIR_CGROUP_CONTROLLER_CPU) |
(1 << VIR_CGROUP_CONTROLLER_CPUACCT) |
(1 << VIR_CGROUP_CONTROLLER_MEMORY) |
vircgroupmock: mock virCgroupV2DevicesAvailable We need to mock virCgroupV2DevicesAvailable() in order to remove any dependency on kernel as BPF devices might not be available. Signed-off-by: Pavel Hrdina <phrdina@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com>
2019-06-24 21:13:37 +08:00
(1 << VIR_CGROUP_CONTROLLER_DEVICES) |
vircgrouptest: add hybrid tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:47 +08:00
(1 << VIR_CGROUP_CONTROLLER_BLKIO);
if (virCgroupNewSelf(&cgroup) < 0) {
fprintf(stderr, "Cannot create cgroup for self\n");
goto cleanup;
}
ret = validateCgroup(cgroup, "", mounts, empty, placement,
"/not/really/sys/fs/cgroup/unified", "/", controllers);
cleanup:
virCgroupFree(&cgroup);
return ret;
}
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
static int testCgroupAvailable(const void *args)
{
bool got = virCgroupAvailable();
bool want = args == (void*)0x1;
if (got != want) {
fprintf(stderr, "Expected cgroup %savailable, but state was wrong\n",
want ? "" : "not ");
return -1;
}
return 0;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupControllerAvailable(const void *args G_GNUC_UNUSED)
vircgroup: Introduce virCgroupControllerAvailable This new internal API checks if given CGroup controller is available. It is going to be needed later when we need to make a decision whether pin domain memory onto NUMA nodes using cpuset CGroup controller or using numa_set_membind(). Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2015-03-31 17:39:13 +08:00
{
int ret = 0;
Remove backslash alignment attempts Right-aligning backslashes when defining macros or using complex commands in Makefiles looks cute, but as soon as any changes is required to the code you end up with either distractingly broken alignment or unnecessarily big diffs where most of the changes are just pushing all backslashes a few characters to one side. Generated using $ git grep -El '[[:blank:]][[:blank:]]\\$' | \ grep -E '*\.([chx]|am|mk)$$' | \ while read f; do \ sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \ done Signed-off-by: Andrea Bolognani <abologna@redhat.com>
2017-11-03 20:09:47 +08:00
# define CHECK_CONTROLLER(c, present) \
if ((present && !virCgroupControllerAvailable(c)) || \
(!present && virCgroupControllerAvailable(c))) { \
fprintf(stderr, present ? \
"Expected controller %s not available\n" : \
"Unexpected controller %s available\n", #c); \
ret = -1; \
vircgroup: Introduce virCgroupControllerAvailable This new internal API checks if given CGroup controller is available. It is going to be needed later when we need to make a decision whether pin domain memory onto NUMA nodes using cpuset CGroup controller or using numa_set_membind(). Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2015-03-31 17:39:13 +08:00
}
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_CPU, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_CPUACCT, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_CPUSET, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_MEMORY, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_DEVICES, false)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_FREEZER, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_BLKIO, true)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_NET_CLS, false)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_PERF_EVENT, false)
CHECK_CONTROLLER(VIR_CGROUP_CONTROLLER_SYSTEMD, true)
# undef CHECK_CONTROLLER
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupGetPercpuStats(const void *args G_GNUC_UNUSED)
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
{
virCgroupPtr cgroup = NULL;
size_t i;
int rv, ret = -1;
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
virTypedParameterPtr params = NULL;
# define EXPECTED_NCPUS 160
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
Fix build with older gcc My commit af1c98e4 broke the build on RHEL-6: vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:566: error: nested extern declaration of '_gl_verify_function2' [-Wnested-externs] The only thing that needs checking is that the array size is at least EXPECTED_NCPUS, to prevent access beyond the array. We can ensure the minimum size also by specifying the array size upfront.
2015-01-23 17:30:01 +08:00
unsigned long long expected[EXPECTED_NCPUS] = {
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
0, 0, 0, 0, 0, 0, 0, 0,
tests: fix build on old 32-bit platforms gcc 4.1.2 (hello RHEL 5) on 32-bit platforms complains: vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:627: warning: integer constant is too large for 'long' type vircgrouptest.c:628: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:634: warning: integer constant is too large for 'long' type vircgrouptest.c:635: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:636: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:644: warning: integer constant is too large for 'long' type * tests/vircgrouptest.c (testCgroupGetPercpuStats): Use ULL suffix. Signed-off-by: Eric Blake <eblake@redhat.com>
2015-04-16 02:50:23 +08:00
7059492996ULL, 0, 0, 0, 0, 0, 0, 0,
4180532496ULL, 0, 0, 0, 0, 0, 0, 0,
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
0, 0, 0, 0, 0, 0, 0, 0,
tests: fix build on old 32-bit platforms gcc 4.1.2 (hello RHEL 5) on 32-bit platforms complains: vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:627: warning: integer constant is too large for 'long' type vircgrouptest.c:628: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:634: warning: integer constant is too large for 'long' type vircgrouptest.c:635: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:636: warning: this decimal constant is unsigned only in ISO C90 vircgrouptest.c:644: warning: integer constant is too large for 'long' type * tests/vircgrouptest.c (testCgroupGetPercpuStats): Use ULL suffix. Signed-off-by: Eric Blake <eblake@redhat.com>
2015-04-16 02:50:23 +08:00
1957541268ULL, 0, 0, 0, 0, 0, 0, 0,
2065932204ULL, 0, 0, 0, 0, 0, 0, 0,
18228689414ULL, 0, 0, 0, 0, 0, 0, 0,
4245525148ULL, 0, 0, 0, 0, 0, 0, 0,
2911161568ULL, 0, 0, 0, 0, 0, 0, 0,
1407758136ULL, 0, 0, 0, 0, 0, 0, 0,
1836807700ULL, 0, 0, 0, 0, 0, 0, 0,
1065296618ULL, 0, 0, 0, 0, 0, 0, 0,
2046213266ULL, 0, 0, 0, 0, 0, 0, 0,
747889778ULL, 0, 0, 0, 0, 0, 0, 0,
709566900ULL, 0, 0, 0, 0, 0, 0, 0,
444777342ULL, 0, 0, 0, 0, 0, 0, 0,
5683512916ULL, 0, 0, 0, 0, 0, 0, 0,
635751356ULL, 0, 0, 0, 0, 0, 0, 0,
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
};
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
if (VIR_ALLOC_N(params, EXPECTED_NCPUS) < 0)
goto cleanup;
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_CPU) |
(1 << VIR_CGROUP_CONTROLLER_CPUACCT),
&cgroup)) < 0) {
fprintf(stderr, "Could not create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
nodeinfo: rename all CPU APIs to have a virHostCPU prefix In preparation for moving all the CPU related APIs out of the nodeinfo file, give them a virHostCPU name prefix. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2016-04-14 01:16:16 +08:00
if (virHostCPUGetCount() != EXPECTED_NCPUS) {
fprintf(stderr, "Unexpected: virHostCPUGetCount() yields: %d\n", virHostCPUGetCount());
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
goto cleanup;
}
if ((rv = virCgroupGetPercpuStats(cgroup,
params,
cgroup: Prepare for sparse vCPU topologies in virCgroupGetPercpuStats Pass a bitmap of enabled guest vCPUs to virCgroupGetPercpuStats so that non-continuous vCPU topologies can be used.
2015-12-14 22:10:22 +08:00
1, 0, EXPECTED_NCPUS, NULL)) < 0) {
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
fprintf(stderr, "Failed call to virCgroupGetPercpuStats for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
for (i = 0; i < EXPECTED_NCPUS; i++) {
Avoid using !STREQ and !STRNEQ We have macros for both positive and negative string matching. Therefore there is no need to use !STREQ or !STRNEQ. At the same time as we are dropping this, new syntax-check rule is introduced to make sure we won't introduce it again. Signed-off-by: Ishmanpreet Kaur Khera <khera.ishman@gmail.com> Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2015-10-21 00:15:12 +08:00
if (STRNEQ(params[i].field, VIR_DOMAIN_CPU_STATS_CPUTIME)) {
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
fprintf(stderr,
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
"Wrong parameter name value from virCgroupGetPercpuStats at %zu (is: %s)\n",
i, params[i].field);
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
goto cleanup;
}
if (params[i].type != VIR_TYPED_PARAM_ULLONG) {
fprintf(stderr,
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
"Wrong parameter value type from virCgroupGetPercpuStats at %zu (is: %d)\n",
i, params[i].type);
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
goto cleanup;
}
if (params[i].value.ul != expected[i]) {
fprintf(stderr,
tests: Use correct function name in error path Commit id 5eb61e6846 neglected to change the name in the wrong value output to virCgroupGetPercpuStats from virCgroupGetMemoryUsage. Signed-off-by: John Ferlan <jferlan@redhat.com>
2018-11-08 07:18:42 +08:00
"Wrong value from virCgroupGetPercpuStats at %zu (expected %llu)\n",
Fix virCgroupGetPercpuStats with non-continuous present CPUs Per-cpu stats are only shown for present CPUs in the cgroups, but we were only parsing the largest CPU number from /sys/devices/system/cpu/present and looking for stats even for non-present CPUs. This resulted in: internal error: cpuacct parse error
2015-01-22 18:54:50 +08:00
i, params[i].value.ul);
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
goto cleanup;
}
}
ret = 0;
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Fix a memory leak in virCgroupGetPercpuStats Coverity reports that my commit af1c98e introduced two memory leaks: the cpumap if ncpus == 0 in virCgroupGetPercpuStats and the params array in the test of the function.
2015-01-26 17:20:22 +08:00
VIR_FREE(params);
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupGetMemoryUsage(const void *args G_GNUC_UNUSED)
Add unit test for virCgroupGetMemoryUsage.
2014-02-15 01:49:06 +08:00
{
virCgroupPtr cgroup = NULL;
int rv, ret = -1;
unsigned long kb;
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_MEMORY),
&cgroup)) < 0) {
fprintf(stderr, "Could not create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupGetMemoryUsage(cgroup, &kb)) < 0) {
fprintf(stderr, "Could not retrieve GetMemoryUsage for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
if (kb != 1421212UL) {
fprintf(stderr,
"Wrong value from virCgroupGetMemoryUsage (expected %ld)\n",
1421212UL);
goto cleanup;
}
ret = 0;
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add unit test for virCgroupGetMemoryUsage.
2014-02-15 01:49:06 +08:00
return ret;
}
tests: Augment vcgrouptest to add virCgroupGetMemoryStat Add a test to fetch the GetMemoryStat output. This only gets data for v1 only right now since the v2 data from commit 61ff6021 is rather useless returning all 0's. The v1 data was originally added in commit d1452470. Signed-off-by: John Ferlan <jferlan@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-11-08 07:40:43 +08:00
static int
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
testCgroupGetMemoryStat(const void *args G_GNUC_UNUSED)
tests: Augment vcgrouptest to add virCgroupGetMemoryStat Add a test to fetch the GetMemoryStat output. This only gets data for v1 only right now since the v2 data from commit 61ff6021 is rather useless returning all 0's. The v1 data was originally added in commit d1452470. Signed-off-by: John Ferlan <jferlan@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-11-08 07:40:43 +08:00
{
virCgroupPtr cgroup = NULL;
int rv;
int ret = -1;
size_t i;
const unsigned long long expected_values[] = {
1336619008ULL,
67100672ULL,
145887232ULL,
661872640ULL,
627400704UL,
3690496ULL
};
const char* names[] = {
"cache",
"active_anon",
"inactive_anon",
"active_file",
"inactive_file",
"unevictable"
};
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
unsigned long long values[G_N_ELEMENTS(expected_values)];
tests: Augment vcgrouptest to add virCgroupGetMemoryStat Add a test to fetch the GetMemoryStat output. This only gets data for v1 only right now since the v2 data from commit 61ff6021 is rather useless returning all 0's. The v1 data was originally added in commit d1452470. Signed-off-by: John Ferlan <jferlan@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-11-08 07:40:43 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_MEMORY),
&cgroup)) < 0) {
fprintf(stderr, "Could not create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupGetMemoryStat(cgroup, &values[0],
&values[1], &values[2],
&values[3], &values[4],
&values[5])) < 0) {
fprintf(stderr, "Could not retrieve GetMemoryStat for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
for (i = 0; i < G_N_ELEMENTS(expected_values); i++) {
tests: Augment vcgrouptest to add virCgroupGetMemoryStat Add a test to fetch the GetMemoryStat output. This only gets data for v1 only right now since the v2 data from commit 61ff6021 is rather useless returning all 0's. The v1 data was originally added in commit d1452470. Signed-off-by: John Ferlan <jferlan@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-11-08 07:40:43 +08:00
/* NB: virCgroupGetMemoryStat returns a KiB scaled value */
if ((expected_values[i] >> 10) != values[i]) {
fprintf(stderr,
"Wrong value (%llu) for %s from virCgroupGetMemoryStat "
"(expected %llu)\n",
values[i], names[i], (expected_values[i] >> 10));
goto cleanup;
}
}
ret = 0;
cleanup:
virCgroupFree(&cgroup);
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupGetBlkioIoServiced(const void *args G_GNUC_UNUSED)
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
{
virCgroupPtr cgroup = NULL;
size_t i;
int rv, ret = -1;
const long long expected_values[] = {
build: fix build on 32-bit hosts vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:543: warning: integer constatnt is too large for 'long' type Signed-off-by: Eric Blake <eblake@redhat.com>
2014-02-22 08:21:58 +08:00
119084214273ULL,
822880960513ULL,
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
9665167,
73283807
};
const char* names[] = {
"bytes read",
"bytes written",
"requests read",
"requests written"
};
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
long long values[G_N_ELEMENTS(expected_values)];
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_BLKIO),
&cgroup)) < 0) {
fprintf(stderr, "Could not create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupGetBlkioIoServiced(cgroup,
values, &values[1],
&values[2], &values[3])) < 0) {
fprintf(stderr, "Could not retrieve BlkioIoServiced for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
for (i = 0; i < G_N_ELEMENTS(expected_values); i++) {
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
if (expected_values[i] != values[i]) {
fprintf(stderr,
"Wrong value for %s from virCgroupBlkioIoServiced (expected %lld)\n",
names[i], expected_values[i]);
goto cleanup;
}
}
ret = 0;
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
return ret;
}
tests: use G_GNUC_UNUSED Use G_GNUC_UNUSED from GLib instead of ATTRIBUTE_UNUSED. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-14 20:45:03 +08:00
static int testCgroupGetBlkioIoDeviceServiced(const void *args G_GNUC_UNUSED)
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
{
virCgroupPtr cgroup = NULL;
size_t i;
int rv, ret = -1;
const long long expected_values0[] = {
build: fix build on 32-bit hosts vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:543: warning: integer constatnt is too large for 'long' type Signed-off-by: Eric Blake <eblake@redhat.com>
2014-02-22 08:21:58 +08:00
59542107136ULL,
411440480256ULL,
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
4832583,
36641903
};
const long long expected_values1[] = {
build: fix build on 32-bit hosts vircgrouptest.c: In function 'testCgroupGetPercpuStats': vircgrouptest.c:543: warning: integer constatnt is too large for 'long' type Signed-off-by: Eric Blake <eblake@redhat.com>
2014-02-22 08:21:58 +08:00
59542107137ULL,
411440480257ULL,
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
4832584,
36641904
};
const char* names[] = {
"bytes read",
"bytes written",
"requests read",
"requests written"
};
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
long long values[G_N_ELEMENTS(expected_values0)];
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
if ((rv = virCgroupNewPartition("/virtualmachines", true,
(1 << VIR_CGROUP_CONTROLLER_BLKIO),
&cgroup)) < 0) {
fprintf(stderr, "Could not create /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
if ((rv = virCgroupGetBlkioIoDeviceServiced(cgroup,
FAKEDEVDIR0,
values, &values[1],
&values[2], &values[3])) < 0) {
fprintf(stderr, "Could not retrieve BlkioIoDeviceServiced for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
for (i = 0; i < G_N_ELEMENTS(expected_values0); i++) {
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
if (expected_values0[i] != values[i]) {
fprintf(stderr,
"Wrong value for %s from virCgroupGetBlkioIoDeviceServiced (expected %lld)\n",
names[i], expected_values0[i]);
goto cleanup;
}
}
if ((rv = virCgroupGetBlkioIoDeviceServiced(cgroup,
FAKEDEVDIR1,
values, &values[1],
&values[2], &values[3])) < 0) {
fprintf(stderr, "Could not retrieve BlkioIoDeviceServiced for /virtualmachines cgroup: %d\n", -rv);
goto cleanup;
}
Use G_N_ELEMENTS instead of ARRAY_CARDINALITY Prefer the GLib version of the macro. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-15 19:55:26 +08:00
for (i = 0; i < G_N_ELEMENTS(expected_values1); i++) {
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
if (expected_values1[i] != values[i]) {
fprintf(stderr,
"Wrong value for %s from virCgroupGetBlkioIoDeviceServiced (expected %lld)\n",
names[i], expected_values1[i]);
goto cleanup;
}
}
ret = 0;
Indent top-level labels by one space in tests/
2014-03-25 14:53:44 +08:00
cleanup:
Revert "util: cgroup: modify virCgroupFree to take virCgroupPtr" This reverts commit 0f80c71822d82465d558d697d3be9af2d21e3675. Turns out, our code relies on virCgroupFree(&var) setting var = NULL. Conflicts: src/util/vircgroup.c: context because 94f1855f099445d is not reverted. Signed-off-by: Michal Privoznik <mprivozn@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-07-30 17:06:31 +08:00
virCgroupFree(&cgroup);
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
return ret;
}
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
tests: Rename LIBVIRT_FAKE_SYSFS_DIR to LIBVIRT_FAKE_ROOT_DIR The old name is no longer accurate, since now we're using its value as the root of the fake filesystem. No functional changes.
2015-12-04 20:38:16 +08:00
# define FAKEROOTDIRTEMPLATE abs_builddir "/fakerootdir-XXXXXX"
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
static char *
initFakeFS(const char *mode,
const char *filename)
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
{
tests: Rename LIBVIRT_FAKE_SYSFS_DIR to LIBVIRT_FAKE_ROOT_DIR The old name is no longer accurate, since now we're using its value as the root of the fake filesystem. No functional changes.
2015-12-04 20:38:16 +08:00
char *fakerootdir;
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
tests: use g_strdup instead of VIR_STRDUP Replace all occurrences of if (VIR_STRDUP(a, b) < 0) /* effectively dead code */ with: a = g_strdup(b); Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Michal Privoznik <mprivozn@redhat.com>
2019-10-20 19:49:46 +08:00
fakerootdir = g_strdup(FAKEROOTDIRTEMPLATE);
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
Use g_mkdtemp instead of mkdtemp Prefer the GLib version to the one from gnulib. Signed-off-by: Ján Tomko <jtomko@redhat.com> Reviewed-by: Peter Krempa <pkrempa@redhat.com>
2019-11-14 05:35:47 +08:00
if (!g_mkdtemp(fakerootdir)) {
tests: Rename LIBVIRT_FAKE_SYSFS_DIR to LIBVIRT_FAKE_ROOT_DIR The old name is no longer accurate, since now we're using its value as the root of the fake filesystem. No functional changes.
2015-12-04 20:38:16 +08:00
fprintf(stderr, "Cannot create fakerootdir");
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
abort();
}
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_setenv("LIBVIRT_FAKE_ROOT_DIR", fakerootdir, TRUE);
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
if (mode)
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_setenv("VIR_CGROUP_MOCK_MODE", mode, TRUE);
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
if (filename)
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_setenv("VIR_CGROUP_MOCK_FILENAME", filename, TRUE);
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
return fakerootdir;
}
static void
cleanupFakeFS(char *fakerootdir)
{
if (getenv("LIBVIRT_SKIP_CLEANUP") == NULL)
virFileDeleteTree(fakerootdir);
VIR_FREE(fakerootdir);
src: switch to use g_setenv/g_unsetenv Eliminate direct use of normal setenv/unsetenv calls in favour of GLib's wrapper. This eliminates two gnulib modules Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Signed-off-by: Daniel P. Berrangé <berrange@redhat.com>
2019-12-19 01:16:19 +08:00
g_unsetenv("LIBVIRT_FAKE_ROOT_DIR");
g_unsetenv("VIR_CGROUP_MOCK_MODE");
g_unsetenv("VIR_CGROUP_MOCK_FILENAME");
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
}
static int
mymain(void)
{
int ret = 0;
char *fakerootdir;
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
# define DETECT_MOUNTS_FULL(file, fail) \
Remove backslash alignment attempts Right-aligning backslashes when defining macros or using complex commands in Makefiles looks cute, but as soon as any changes is required to the code you end up with either distractingly broken alignment or unnecessarily big diffs where most of the changes are just pushing all backslashes a few characters to one side. Generated using $ git grep -El '[[:blank:]][[:blank:]]\\$' | \ grep -E '*\.([chx]|am|mk)$$' | \ while read f; do \ sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \ done Signed-off-by: Andrea Bolognani <abologna@redhat.com>
2017-11-03 20:09:47 +08:00
do { \
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
struct _detectMountsData data = { file, fail }; \
Remove backslash alignment attempts Right-aligning backslashes when defining macros or using complex commands in Makefiles looks cute, but as soon as any changes is required to the code you end up with either distractingly broken alignment or unnecessarily big diffs where most of the changes are just pushing all backslashes a few characters to one side. Generated using $ git grep -El '[[:blank:]][[:blank:]]\\$' | \ grep -E '*\.([chx]|am|mk)$$' | \ while read f; do \ sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \ done Signed-off-by: Andrea Bolognani <abologna@redhat.com>
2017-11-03 20:09:47 +08:00
if (virTestRun("Detect cgroup mounts for " file, \
testCgroupDetectMounts, \
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
&data) < 0) \
Remove backslash alignment attempts Right-aligning backslashes when defining macros or using complex commands in Makefiles looks cute, but as soon as any changes is required to the code you end up with either distractingly broken alignment or unnecessarily big diffs where most of the changes are just pushing all backslashes a few characters to one side. Generated using $ git grep -El '[[:blank:]][[:blank:]]\\$' | \ grep -E '*\.([chx]|am|mk)$$' | \ while read f; do \ sed -Ei 's/[[:blank:]]*[[:blank:]]\\$/ \\/g' "$f"; \ done Signed-off-by: Andrea Bolognani <abologna@redhat.com>
2017-11-03 20:09:47 +08:00
ret = -1; \
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
} while (0)
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
# define DETECT_MOUNTS(file) DETECT_MOUNTS_FULL(file, false);
# define DETECT_MOUNTS_FAIL(file) DETECT_MOUNTS_FULL(file, true);
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
DETECT_MOUNTS("ovirt-node-6.6");
DETECT_MOUNTS("ovirt-node-7.1");
DETECT_MOUNTS("fedora-18");
DETECT_MOUNTS("fedora-21");
DETECT_MOUNTS("rhel-7.1");
DETECT_MOUNTS("cgroups1");
DETECT_MOUNTS("cgroups2");
DETECT_MOUNTS("cgroups3");
DETECT_MOUNTS("all-in-one");
vircgrouptest: call virCgroupNewSelf instead virCgroupDetectMounts This will be required once cgroup v2 is introduced. The cgroup detection is not simple and we will have multiple backends so we should not just jump into the middle of the detection code. In order to use virCgroupNewSelf we need to create all the remaining data files: - {name}.cgroups represents /proc/cgroups, it is a list of cgroup controllers compiled into kernel - {name}.self.cgroup represents /proc/self/cgroup, it describes cgroups to which the process belongs For "no-cgroups" we need to modify the expected behavior because virCgroupNewSelf() will fail if there are no controllers available. Reviewed-by: Fabiano Fidêncio <fidencio@redhat.com> Reviewed-by: Ján Tomko <jtomko@redhat.com> Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-14 19:17:07 +08:00
DETECT_MOUNTS_FAIL("no-cgroups");
Avoid hidden cgroup mount points Currently the scan of the /proc/mounts file used to find cgroup mount points doesn't take into account that mount points may hidden by other mount points. For, example in certain Kubernetes environments the /proc/mounts contains the following lines: cgroup /sys/fs/cgroup/net_prio,net_cls cgroup ... tmpfs /sys/fs/cgroup tmpfs ... cgroup /sys/fs/cgroup/net_cls,net_prio cgroup ... In this particular environment the first mount point is hidden by the second one. The correct mount point is the third one, but libvirt will never process it because it only checks the first mount point for each controller (net_cls in this case). So libvirt will try to use the first mount point, which doesn't actually exist, and the complete detection process will fail. To avoid that issue this patch changes the virCgroupDetectMountsFromFile function so that when there are duplicates it takes the information from the last line in /proc/mounts. This requires removing the previous explicit condition to skip duplicates, and adding code to free the memory used by the processing of duplicated lines. Related-To: https://bugzilla.redhat.com/1468214 Related-To: https://github.com/kubevirt/libvirt/issues/4 Signed-off-by: Juan Hernandez <jhernand@redhat.com>
2017-07-06 23:03:31 +08:00
DETECT_MOUNTS("kubevirt");
vircgrouptest: add detect mounts test for cgroup v2 Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 16:08:35 +08:00
fakerootdir = initFakeFS("unified", NULL);
DETECT_MOUNTS("unified");
cleanupFakeFS(fakerootdir);
vircgrouptest: add detect mounts test for hybrid cgroups Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 16:09:20 +08:00
fakerootdir = initFakeFS("hybrid", NULL);
DETECT_MOUNTS("hybrid");
cleanupFakeFS(fakerootdir);
tests: Add tests for virCgroupDetectMounts Signed-off-by: Jiri Denemark <jdenemar@redhat.com>
2015-03-17 05:36:52 +08:00
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
fakerootdir = initFakeFS(NULL, "systemd");
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for self", testCgroupNewForSelf, NULL) < 0)
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for partition", testCgroupNewForPartition, NULL) < 0)
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for partition nested", testCgroupNewForPartitionNested, NULL) < 0)
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for partition nested deeply", testCgroupNewForPartitionNestedDeep, NULL) < 0)
Ensure all cgroup partitions have a suffix of ".partition" If the partition named passed in the XML does not already have a suffix, ensure it gets a '.partition' added to each component. The exceptions are /machine, /user and /system which do not need to have a suffix, since they are fixed partitions at the top level. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:23:51 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for domain partition", testCgroupNewForPartitionDomain, NULL) < 0)
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for domain partition escaped", testCgroupNewForPartitionDomainEscaped, NULL) < 0)
Do proper escaping of cgroup resource partitions If a user cgroup name begins with "cgroup.", "_" or with any of the controllers from /proc/cgroups followed by a dot, then they need to be prefixed with a single underscore. eg if there is an object "cpu.service", then this would end up as "_cpu.service" in the cgroup filesystem tree, however, "waldo.service" would stay "waldo.service", at least as long as nobody comes up with a cgroup controller called "waldo". Since we require a '.XXXX' suffix on all partitions, there is no scope for clashing with the kernel 'tasks' and 'release_agent' files. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-26 17:50:24 +08:00
ret = -1;
Add a new virCgroupNewPartition for setting up resource partitions A resource partition is an absolute cgroup path, ignoring the current process placement. Expose a virCgroupNewPartition API for constructing such cgroups Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-29 02:08:39 +08:00
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("Cgroup available", testCgroupAvailable, (void*)0x1) < 0)
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("Cgroup controller available", testCgroupControllerAvailable, NULL) < 0)
vircgroup: Introduce virCgroupControllerAvailable This new internal API checks if given CGroup controller is available. It is going to be needed later when we need to make a decision whether pin domain memory onto NUMA nodes using cpuset CGroup controller or using numa_set_membind(). Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2015-03-31 17:39:13 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("virCgroupGetBlkioIoServiced works", testCgroupGetBlkioIoServiced, NULL) < 0)
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("virCgroupGetBlkioIoDeviceServiced works", testCgroupGetBlkioIoDeviceServiced, NULL) < 0)
Add unit test for virCgroupGetBlkioIo*Serviced
2014-02-15 01:49:05 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("virCgroupGetMemoryUsage works", testCgroupGetMemoryUsage, NULL) < 0)
Add unit test for virCgroupGetMemoryUsage.
2014-02-15 01:49:06 +08:00
ret = -1;
tests: Augment vcgrouptest to add virCgroupGetMemoryStat Add a test to fetch the GetMemoryStat output. This only gets data for v1 only right now since the v2 data from commit 61ff6021 is rather useless returning all 0's. The v1 data was originally added in commit d1452470. Signed-off-by: John Ferlan <jferlan@redhat.com> Reviewed-by: Pavel Hrdina <phrdina@redhat.com>
2018-11-08 07:40:43 +08:00
if (virTestRun("virCgroupGetMemoryStat works", testCgroupGetMemoryStat, NULL) < 0)
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("virCgroupGetPercpuStats works", testCgroupGetPercpuStats, NULL) < 0)
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
ret = -1;
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
cleanupFakeFS(fakerootdir);
Add unit test for virCgroupGetPercpuStats.
2014-02-15 01:49:08 +08:00
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
fakerootdir = initFakeFS(NULL, "all-in-one");
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for self (allinone)", testCgroupNewForSelfAllInOne, NULL) < 0)
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("Cgroup available", testCgroupAvailable, (void*)0x1) < 0)
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
ret = -1;
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
cleanupFakeFS(fakerootdir);
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
fakerootdir = initFakeFS(NULL, "logind");
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("New cgroup for self (logind)", testCgroupNewForSelfLogind, NULL) < 0)
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
ret = -1;
tests: Rename virtTestRun to virTestRun. This function doesn't follow our convention of naming functions.
2016-05-26 23:01:50 +08:00
if (virTestRun("Cgroup available", testCgroupAvailable, (void*)0x0) < 0)
Fix launching of VMs on when only logind part of systemd is present Debian systems may run the 'systemd-logind' daemon, which causes the /sys/fs/cgroup/systemd mount to be setup, but no other cgroup controllers are created. While the LXC driver considers cgroups to be mandatory, the QEMU driver is supposed to accept them as optional. We detect whether they are present by looking in /proc/mounts for any mounts of type 'cgroups', but this is not sufficient. We need to skip any named mounts (as seen by a name=XXX string in the mount options), so that we only detect actual resource controllers. http://bugs.debian.org/cgi-bin/bugreport.cgi?bug=721979 Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-12 02:15:52 +08:00
ret = -1;
vircgrouptest: introduce initFakeFS and cleanupFakeFS helpers We need to configure multiple env variables for each set of tests so create helper functions to do that. Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-24 18:22:54 +08:00
cleanupFakeFS(fakerootdir);
Fix cgroups when all are mounted on /sys/fs/cgroup Some users in Ubuntu/Debian seem to have a setup where all the cgroup controllers are mounted on /sys/fs/cgroup rather than any /sys/fs/cgroup/<controller> name. In the loop which detects which controllers are present for a mount point we were modifying 'mnt_dir' field in the 'struct mntent' var, but not always restoring the original value. This caused detection to break in the all-in-one mount setup. Fix that logic bug and add test case coverage for this mount setup. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-09-10 21:31:53 +08:00
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
/* cgroup unified */
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
vircgrouptest: add cgroup v2 tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:09 +08:00
fakerootdir = initFakeFS("unified", "unified");
if (virTestRun("New cgroup for self (unified)", testCgroupNewForSelfUnified, NULL) < 0)
ret = -1;
if (virTestRun("Cgroup available (unified)", testCgroupAvailable, (void*)0x1) < 0)
ret = -1;
cleanupFakeFS(fakerootdir);
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
vircgrouptest: add hybrid tests Signed-off-by: Pavel Hrdina <phrdina@redhat.com>
2018-09-18 15:43:47 +08:00
/* cgroup hybrid */
fakerootdir = initFakeFS("hybrid", "hybrid");
if (virTestRun("New cgroup for self (hybrid)", testCgroupNewForSelfHybrid, NULL) < 0)
ret = -1;
if (virTestRun("Cgroup available (hybrid)", testCgroupAvailable, (void*)0x1) < 0)
ret = -1;
cleanupFakeFS(fakerootdir);
Require spaces around equality comparisons Commit a1cbe4b5 added a check for spaces around assignments and this patch extends it to checks for spaces around '=='. One exception is virAssertCmpInt where comma after '==' is acceptable (since it is a macro and '==' is its argument). Signed-off-by: Martin Kletzander <mkletzan@redhat.com>
2014-03-17 17:38:38 +08:00
return ret == 0 ? EXIT_SUCCESS : EXIT_FAILURE;
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
}
tests: Add lib- prefix to all mocks In preparation libtool "-module" flag removal, add lib prefix to all mock shared objects. While at it, introduce VIR_TEST_MOCK macros that makes path out of mock name to be used with VIR_TEST_PRELOAD or VIR_TEST_MAIN_PRELOAD. That, hopefully, improves readability, reduces line length and allows to tailor VIR_TEST_MOCK for specific platform if it has shared library suffix different from ".so". Signed-off-by: Roman Bolshakov <r.bolshakov@yadro.com>
2019-08-22 00:13:16 +08:00
VIR_TEST_MAIN_PRELOAD(mymain, VIR_TEST_MOCK("vircgroup"))
Add a test suite for cgroups functionality Some aspects of the cgroups setup / detection code are quite subtle and easy to break. It would greatly benefit from unit testing, but this is difficult because the test suite won't have privileges to play around with cgroups. The solution is to use monkey patching via LD_PRELOAD to override the fopen, open, mkdir, access functions to redirect access of cgroups files to some magic stubs in the test suite. Using this we provide custom content for the /proc/cgroup and /proc/self/mounts files which report a fixed cgroup setup. We then override open/mkdir/access so that access to the cgroups filesystem gets redirected into files in a temporary directory tree in the test suite build dir. Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-03-28 22:36:52 +08:00
#else
int
main(void)
{
return EXIT_AM_SKIP;
}
#endif