From: "Daniel P. Berrange" <berrange@xxxxxxxxxx>
To bring in line with new naming practice, rename the=
src/util/cgroup.{h,c} files to vircgroup.{h,c}
Signed-off-by: Daniel P. Berrange <berrange@xxxxxxxxxx>
---
po/POTFILES.in | 2 +-
src/Makefile.am | 2 +-
src/conf/domain_audit.h | 2 +-
src/lxc/lxc_cgroup.c | 2 +-
src/lxc/lxc_conf.h | 2 +-
src/qemu/qemu_cgroup.c | 2 +-
src/qemu/qemu_conf.h | 2 +-
src/util/cgroup.c | 2099 -----------------------------------------------
src/util/cgroup.h | 167 ----
src/util/vircgroup.c | 2099 +++++++++++++++++++++++++++++++++++++++++++++++
src/util/vircgroup.h | 167 ++++
11 files changed, 2273 insertions(+), 2273 deletions(-)
delete mode 100644 src/util/cgroup.c
delete mode 100644 src/util/cgroup.h
create mode 100644 src/util/vircgroup.c
create mode 100644 src/util/vircgroup.h
diff --git a/po/POTFILES.in b/po/POTFILES.in
index f0cfd7f..19ed187 100644
--- a/po/POTFILES.in
+++ b/po/POTFILES.in
@@ -137,7 +137,6 @@ src/storage/storage_driver.c
src/test/test_driver.c
src/uml/uml_conf.c
src/uml/uml_driver.c
-src/util/cgroup.c
src/util/command.c
src/util/conf.c
src/util/dnsmasq.c
@@ -157,6 +156,7 @@ src/util/util.c
src/util/viraudit.c
src/util/virauth.c
src/util/virauthconfig.c
+src/util/vircgroup.c
src/util/virdbus.c
src/util/virfile.c
src/util/virhash.c
diff --git a/src/Makefile.am b/src/Makefile.am
index 53ec9da..149ffc9 100644
--- a/src/Makefile.am
+++ b/src/Makefile.am
@@ -57,7 +57,6 @@ UTIL_SOURCES = \
util/buf.c util/buf.h \
util/command.c util/command.h \
util/conf.c util/conf.h \
- util/cgroup.c util/cgroup.h \
util/event.c util/event.h \
util/event_poll.c util/event_poll.h \
util/hooks.c util/hooks.h \
@@ -92,6 +91,7 @@ UTIL_SOURCES = \
util/virtypedparam.c util/virtypedparam.h \
util/xml.c util/xml.h \
util/virterror.c util/virterror_internal.h \
+ util/vircgroup.c util/vircgroup.h \
util/virdbus.c util/virdbus.h \
util/virhash.c util/virhash.h \
util/virhashcode.c util/virhashcode.h \
diff --git a/src/conf/domain_audit.h b/src/conf/domain_audit.h
index 63a8f75..381fe37 100644
--- a/src/conf/domain_audit.h
+++ b/src/conf/domain_audit.h
@@ -25,7 +25,7 @@
# define __VIR_DOMAIN_AUDIT_H__
# include "domain_conf.h"
-# include "cgroup.h"
+# include "vircgroup.h"
void virDomainAuditStart(virDomainObjPtr vm,
const char *reason,
diff --git a/src/lxc/lxc_cgroup.c b/src/lxc/lxc_cgroup.c
index 767ef26..e354d6c 100644
--- a/src/lxc/lxc_cgroup.c
+++ b/src/lxc/lxc_cgroup.c
@@ -27,7 +27,7 @@
#include "virterror_internal.h"
#include "logging.h"
#include "memory.h"
-#include "cgroup.h"
+#include "vircgroup.h"
#define VIR_FROM_THIS VIR_FROM_LXC
diff --git a/src/lxc/lxc_conf.h b/src/lxc/lxc_conf.h
index ea345be..c02966f 100644
--- a/src/lxc/lxc_conf.h
+++ b/src/lxc/lxc_conf.h
@@ -32,7 +32,7 @@
# include "domain_event.h"
# include "capabilities.h"
# include "threads.h"
-# include "cgroup.h"
+# include "vircgroup.h"
# include "security/security_manager.h"
# include "configmake.h"
diff --git a/src/qemu/qemu_cgroup.c b/src/qemu/qemu_cgroup.c
index 30cd1d6..6c6607d 100644
--- a/src/qemu/qemu_cgroup.c
+++ b/src/qemu/qemu_cgroup.c
@@ -26,7 +26,7 @@
#include "qemu_cgroup.h"
#include "qemu_domain.h"
#include "qemu_process.h"
-#include "cgroup.h"
+#include "vircgroup.h"
#include "logging.h"
#include "memory.h"
#include "virterror_internal.h"
diff --git a/src/qemu/qemu_conf.h b/src/qemu/qemu_conf.h
index 1a39946..cfa6fff 100644
--- a/src/qemu/qemu_conf.h
+++ b/src/qemu/qemu_conf.h
@@ -34,7 +34,7 @@
# include "domain_event.h"
# include "threads.h"
# include "security/security_manager.h"
-# include "cgroup.h"
+# include "vircgroup.h"
# include "pci.h"
# include "hostusb.h"
# include "cpu_conf.h"
diff --git a/src/util/cgroup.c b/src/util/cgroup.c
deleted file mode 100644
index 8f3b8b7..0000000
--- a/src/util/cgroup.c
+++ /dev/null
@@ -1,2099 +0,0 @@
-/*
- * cgroup.c: Tools for managing cgroups
- *
- * Copyright (C) 2010-2012 Red Hat, Inc.
- * Copyright IBM Corp. 2008
- *
- * 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/>.
- *
- * Authors:
- * Dan Smith <danms@xxxxxxxxxx>
- */
-#include <config.h>
-
-#include <stdio.h>
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-# include <mntent.h>
-#endif
-#include <fcntl.h>
-#include <string.h>
-#include <errno.h>
-#include <stdlib.h>
-#include <sys/stat.h>
-#include <sys/types.h>
-#include <signal.h>
-#include <libgen.h>
-#include <dirent.h>
-
-#include "internal.h"
-#include "util.h"
-#include "memory.h"
-#include "cgroup.h"
-#include "logging.h"
-#include "virfile.h"
-#include "virhash.h"
-#include "virhashcode.h"
-
-#define CGROUP_MAX_VAL 512
-
-VIR_ENUM_IMPL(virCgroupController, VIR_CGROUP_CONTROLLER_LAST,
- "cpu", "cpuacct", "cpuset", "memory", "devices",
- "freezer", "blkio");
-
-struct virCgroupController {
- int type;
- char *mountPoint;
- char *placement;
-};
-
-struct virCgroup {
- char *path;
-
- struct virCgroupController controllers[VIR_CGROUP_CONTROLLER_LAST];
-};
-
-typedef enum {
- VIR_CGROUP_NONE = 0, /* create subdir under each cgroup if possible. */
- VIR_CGROUP_MEM_HIERACHY = 1 << 0, /* call virCgroupSetMemoryUseHierarchy
- * before creating subcgroups and
- * attaching tasks
- */
- VIR_CGROUP_VCPU = 1 << 1, /* create subdir only under the cgroup cpu,
- * cpuacct and cpuset if possible. */
-} virCgroupFlags;
-
-/**
- * virCgroupFree:
- *
- * @group: The group structure to free
- */
-void virCgroupFree(virCgroupPtr *group)
-{
- int i;
-
- if (*group == NULL)
- return;
-
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- VIR_FREE((*group)->controllers[i].mountPoint);
- VIR_FREE((*group)->controllers[i].placement);
- }
-
- VIR_FREE((*group)->path);
- VIR_FREE(*group);
-}
-
-/**
- * virCgroupMounted: query whether a cgroup subsystem is mounted or not
- *
- * @cgroup: The group structure to be queried
- * @controller: cgroup subsystem id
- *
- * Returns true if a cgroup is subsystem is mounted.
- */
-bool virCgroupMounted(virCgroupPtr cgroup, int controller)
-{
- return cgroup->controllers[controller].mountPoint != NULL;
-}
-
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-/*
- * Process /proc/mounts figuring out what controllers are
- * mounted and where
- */
-static int virCgroupDetectMounts(virCgroupPtr group)
-{
- int i;
- FILE *mounts = NULL;
- struct mntent entry;
- char buf[CGROUP_MAX_VAL];
-
- mounts = fopen("/proc/mounts", "r");
- if (mounts == NULL) {
- VIR_ERROR(_("Unable to open /proc/mounts"));
- return -ENOENT;
- }
-
- while (getmntent_r(mounts, &entry, buf, sizeof(buf)) != NULL) {
- if (STRNEQ(entry.mnt_type, "cgroup"))
- continue;
-
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- const char *typestr = virCgroupControllerTypeToString(i);
- int typelen = strlen(typestr);
- char *tmp = entry.mnt_opts;
- while (tmp) {
- char *next = strchr(tmp, ',');
- int len;
- if (next) {
- len = next-tmp;
- next++;
- } else {
- len = strlen(tmp);
- }
- /* NB, the same controller can appear >1 time in mount list
- * due to bind mounts from one location to another. Pick the
- * first entry only
- */
- if (typelen == len && STREQLEN(typestr, tmp, len) &&
- !group->controllers[i].mountPoint &&
- !(group->controllers[i].mountPoint = strdup(entry.mnt_dir)))
- goto no_memory;
- tmp = next;
- }
- }
- }
-
- VIR_FORCE_FCLOSE(mounts);
-
- return 0;
-
-no_memory:
- VIR_FORCE_FCLOSE(mounts);
- return -ENOMEM;
-}
-
-
-/*
- * Process /proc/self/cgroup figuring out what cgroup
- * sub-path the current process is assigned to. ie not
- * necessarily in the root
- */
-static int virCgroupDetectPlacement(virCgroupPtr group)
-{
- int i;
- FILE *mapping = NULL;
- char line[1024];
-
- mapping = fopen("/proc/self/cgroup", "r");
- if (mapping == NULL) {
- VIR_ERROR(_("Unable to open /proc/self/cgroup"));
- return -ENOENT;
- }
-
- while (fgets(line, sizeof(line), mapping) != NULL) {
- char *controllers = strchr(line, ':');
- char *path = controllers ? strchr(controllers+1, ':') : NULL;
- char *nl = path ? strchr(path, '\n') : NULL;
-
- if (!controllers || !path)
- continue;
-
- if (nl)
- *nl = '\0';
-
- *path = '\0';
- controllers++;
- path++;
-
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- const char *typestr = virCgroupControllerTypeToString(i);
- int typelen = strlen(typestr);
- char *tmp = controllers;
- while (tmp) {
- char *next = strchr(tmp, ',');
- int len;
- if (next) {
- len = next-tmp;
- next++;
- } else {
- len = strlen(tmp);
- }
- if (typelen == len && STREQLEN(typestr, tmp, len) &&
- !(group->controllers[i].placement = strdup(STREQ(path, "/") ? "" : path)))
- goto no_memory;
-
- tmp = next;
- }
- }
- }
-
- VIR_FORCE_FCLOSE(mapping);
-
- return 0;
-
-no_memory:
- VIR_FORCE_FCLOSE(mapping);
- return -ENOMEM;
-
-}
-
-static int virCgroupDetect(virCgroupPtr group)
-{
- int any = 0;
- int rc;
- int i;
-
- rc = virCgroupDetectMounts(group);
- if (rc < 0) {
- VIR_ERROR(_("Failed to detect mounts for %s"), group->path);
- return rc;
- }
-
- /* Check that at least 1 controller is available */
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- if (group->controllers[i].mountPoint != NULL)
- any = 1;
- }
- if (!any)
- return -ENXIO;
-
-
- rc = virCgroupDetectPlacement(group);
-
- if (rc == 0) {
- /* Check that for every mounted controller, we found our placement */
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- if (!group->controllers[i].mountPoint)
- continue;
-
- if (!group->controllers[i].placement) {
- VIR_ERROR(_("Could not find placement for controller %s at %s"),
- virCgroupControllerTypeToString(i),
- group->controllers[i].placement);
- rc = -ENOENT;
- break;
- }
-
- VIR_DEBUG("Detected mount/mapping %i:%s at %s in %s", i,
- virCgroupControllerTypeToString(i),
- group->controllers[i].mountPoint,
- group->controllers[i].placement);
- }
- } else {
- VIR_ERROR(_("Failed to detect mapping for %s"), group->path);
- }
-
- return rc;
-}
-#endif
-
-
-int virCgroupPathOfController(virCgroupPtr group,
- int controller,
- const char *key,
- char **path)
-{
- if (controller == -1) {
- int i;
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- if (group->controllers[i].mountPoint &&
- group->controllers[i].placement) {
- controller = i;
- break;
- }
- }
- }
- if (controller == -1)
- return -ENOSYS;
-
- if (group->controllers[controller].mountPoint == NULL)
- return -ENOENT;
-
- if (group->controllers[controller].placement == NULL)
- return -ENOENT;
-
- if (virAsprintf(path, "%s%s%s/%s",
- group->controllers[controller].mountPoint,
- group->controllers[controller].placement,
- STREQ(group->path, "/") ? "" : group->path,
- key ? key : "") == -1)
- return -ENOMEM;
-
- return 0;
-}
-
-
-static int virCgroupSetValueStr(virCgroupPtr group,
- int controller,
- const char *key,
- const char *value)
-{
- int rc = 0;
- char *keypath = NULL;
-
- rc = virCgroupPathOfController(group, controller, key, &keypath);
- if (rc != 0)
- return rc;
-
- VIR_DEBUG("Set value '%s' to '%s'", keypath, value);
- rc = virFileWriteStr(keypath, value, 0);
- if (rc < 0) {
- rc = -errno;
- VIR_DEBUG("Failed to write value '%s': %m", value);
- } else {
- rc = 0;
- }
-
- VIR_FREE(keypath);
-
- return rc;
-}
-
-static int virCgroupGetValueStr(virCgroupPtr group,
- int controller,
- const char *key,
- char **value)
-{
- int rc;
- char *keypath = NULL;
-
- *value = NULL;
-
- rc = virCgroupPathOfController(group, controller, key, &keypath);
- if (rc != 0) {
- VIR_DEBUG("No path of %s, %s", group->path, key);
- return rc;
- }
-
- VIR_DEBUG("Get value %s", keypath);
-
- rc = virFileReadAll(keypath, 1024*1024, value);
- if (rc < 0) {
- rc = -errno;
- VIR_DEBUG("Failed to read %s: %m\n", keypath);
- } else {
- /* Terminated with '\n' has sometimes harmful effects to the caller */
- if ((*value)[rc - 1] == '\n')
- (*value)[rc - 1] = '\0';
-
- rc = 0;
- }
-
- VIR_FREE(keypath);
-
- return rc;
-}
-
-static int virCgroupSetValueU64(virCgroupPtr group,
- int controller,
- const char *key,
- unsigned long long int value)
-{
- char *strval = NULL;
- int rc;
-
- if (virAsprintf(&strval, "%llu", value) == -1)
- return -ENOMEM;
-
- rc = virCgroupSetValueStr(group, controller, key, strval);
-
- VIR_FREE(strval);
-
- return rc;
-}
-
-
-
-static int virCgroupSetValueI64(virCgroupPtr group,
- int controller,
- const char *key,
- long long int value)
-{
- char *strval = NULL;
- int rc;
-
- if (virAsprintf(&strval, "%lld", value) == -1)
- return -ENOMEM;
-
- rc = virCgroupSetValueStr(group, controller, key, strval);
-
- VIR_FREE(strval);
-
- return rc;
-}
-
-static int virCgroupGetValueI64(virCgroupPtr group,
- int controller,
- const char *key,
- long long int *value)
-{
- char *strval = NULL;
- int rc = 0;
-
- rc = virCgroupGetValueStr(group, controller, key, &strval);
- if (rc != 0)
- goto out;
-
- if (virStrToLong_ll(strval, NULL, 10, value) < 0)
- rc = -EINVAL;
-out:
- VIR_FREE(strval);
-
- return rc;
-}
-
-static int virCgroupGetValueU64(virCgroupPtr group,
- int controller,
- const char *key,
- unsigned long long int *value)
-{
- char *strval = NULL;
- int rc = 0;
-
- rc = virCgroupGetValueStr(group, controller, key, &strval);
- if (rc != 0)
- goto out;
-
- if (virStrToLong_ull(strval, NULL, 10, value) < 0)
- rc = -EINVAL;
-out:
- VIR_FREE(strval);
-
- return rc;
-}
-
-
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-static int virCgroupCpuSetInherit(virCgroupPtr parent, virCgroupPtr group)
-{
- int i;
- int rc = 0;
- const char *inherit_values[] = {
- "cpuset.cpus",
- "cpuset.mems",
- };
-
- VIR_DEBUG("Setting up inheritance %s -> %s", parent->path, group->path);
- for (i = 0; i < ARRAY_CARDINALITY(inherit_values) ; i++) {
- char *value;
-
- rc = virCgroupGetValueStr(parent,
- VIR_CGROUP_CONTROLLER_CPUSET,
- inherit_values[i],
- &value);
- if (rc != 0) {
- VIR_ERROR(_("Failed to get %s %d"), inherit_values[i], rc);
- break;
- }
-
- VIR_DEBUG("Inherit %s = %s", inherit_values[i], value);
-
- rc = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_CPUSET,
- inherit_values[i],
- value);
- VIR_FREE(value);
-
- if (rc != 0) {
- VIR_ERROR(_("Failed to set %s %d"), inherit_values[i], rc);
- break;
- }
- }
-
- return rc;
-}
-
-static int virCgroupSetMemoryUseHierarchy(virCgroupPtr group)
-{
- int rc = 0;
- unsigned long long value;
- const char *filename = "memory.use_hierarchy";
-
- rc = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- filename, &value);
- if (rc != 0) {
- VIR_ERROR(_("Failed to read %s/%s (%d)"), group->path, filename, rc);
- return rc;
- }
-
- /* Setting twice causes error, so if already enabled, skip setting */
- if (value == 1)
- return 0;
-
- VIR_DEBUG("Setting up %s/%s", group->path, filename);
- rc = virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- filename, 1);
-
- if (rc != 0) {
- VIR_ERROR(_("Failed to set %s/%s (%d)"), group->path, filename, rc);
- }
-
- return rc;
-}
-
-static int virCgroupMakeGroup(virCgroupPtr parent,
- virCgroupPtr group,
- bool create,
- unsigned int flags)
-{
- int i;
- int rc = 0;
-
- VIR_DEBUG("Make group %s", group->path);
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- char *path = NULL;
-
- /* Skip over controllers that aren't mounted */
- if (!group->controllers[i].mountPoint)
- continue;
-
- /* We need to control cpu bandwidth for each vcpu now */
- if ((flags & VIR_CGROUP_VCPU) &&
- (i != VIR_CGROUP_CONTROLLER_CPU &&
- i != VIR_CGROUP_CONTROLLER_CPUACCT &&
- i != VIR_CGROUP_CONTROLLER_CPUSET)) {
- /* treat it as unmounted and we can use virCgroupAddTask */
- VIR_FREE(group->controllers[i].mountPoint);
- continue;
- }
-
- rc = virCgroupPathOfController(group, i, "", &path);
- if (rc < 0)
- return rc;
- /* As of Feb 2011, clang can't see that the above function
- * call did not modify group. */
- sa_assert(group->controllers[i].mountPoint);
-
- VIR_DEBUG("Make controller %s", path);
- if (access(path, F_OK) != 0) {
- if (!create ||
- mkdir(path, 0755) < 0) {
- /* With a kernel that doesn't support multi-level directory
- * for blkio controller, libvirt will fail and disable all
- * other controllers even though they are available. So
- * treat blkio as unmounted if mkdir fails. */
- if (i == VIR_CGROUP_CONTROLLER_BLKIO) {
- rc = 0;
- VIR_FREE(group->controllers[i].mountPoint);
- VIR_FREE(path);
- continue;
- } else {
- rc = -errno;
- VIR_FREE(path);
- break;
- }
- }
- if (group->controllers[VIR_CGROUP_CONTROLLER_CPUSET].mountPoint != NULL &&
- (i == VIR_CGROUP_CONTROLLER_CPUSET ||
- STREQ(group->controllers[i].mountPoint, group->controllers[VIR_CGROUP_CONTROLLER_CPUSET].mountPoint))) {
- rc = virCgroupCpuSetInherit(parent, group);
- if (rc != 0) {
- VIR_FREE(path);
- break;
- }
- }
- /*
- * Note that virCgroupSetMemoryUseHierarchy should always be
- * called prior to creating subcgroups and attaching tasks.
- */
- if ((flags & VIR_CGROUP_MEM_HIERACHY) &&
- (group->controllers[VIR_CGROUP_CONTROLLER_MEMORY].mountPoint != NULL) &&
- (i == VIR_CGROUP_CONTROLLER_MEMORY ||
- STREQ(group->controllers[i].mountPoint, group->controllers[VIR_CGROUP_CONTROLLER_MEMORY].mountPoint))) {
- rc = virCgroupSetMemoryUseHierarchy(group);
- if (rc != 0) {
- VIR_FREE(path);
- break;
- }
- }
- }
-
- VIR_FREE(path);
- }
-
- return rc;
-}
-
-
-static int virCgroupNew(const char *path,
- virCgroupPtr *group)
-{
- int rc = 0;
- char *typpath = NULL;
-
- VIR_DEBUG("New group %s", path);
- *group = NULL;
-
- if (VIR_ALLOC((*group)) != 0) {
- rc = -ENOMEM;
- goto err;
- }
-
- if (!((*group)->path = strdup(path))) {
- rc = -ENOMEM;
- goto err;
- }
-
- rc = virCgroupDetect(*group);
- if (rc < 0)
- goto err;
-
- return rc;
-err:
- virCgroupFree(group);
- *group = NULL;
-
- VIR_FREE(typpath);
-
- return rc;
-}
-
-static int virCgroupAppRoot(bool privileged,
- virCgroupPtr *group,
- bool create)
-{
- virCgroupPtr rootgrp = NULL;
- int rc;
-
- rc = virCgroupNew("/", &rootgrp);
- if (rc != 0)
- return rc;
-
- if (privileged) {
- rc = virCgroupNew("/libvirt", group);
- } else {
- char *rootname;
- char *username;
- username = virGetUserName(getuid());
- if (!username) {
- rc = -ENOMEM;
- goto cleanup;
- }
- rc = virAsprintf(&rootname, "/libvirt-%s", username);
- VIR_FREE(username);
- if (rc < 0) {
- rc = -ENOMEM;
- goto cleanup;
- }
-
- rc = virCgroupNew(rootname, group);
- VIR_FREE(rootname);
- }
- if (rc != 0)
- goto cleanup;
-
- rc = virCgroupMakeGroup(rootgrp, *group, create, VIR_CGROUP_NONE);
-
-cleanup:
- virCgroupFree(&rootgrp);
- return rc;
-}
-#endif
-
-#if defined _DIRENT_HAVE_D_TYPE
-static int virCgroupRemoveRecursively(char *grppath)
-{
- DIR *grpdir;
- struct dirent *ent;
- int rc = 0;
-
- grpdir = opendir(grppath);
- if (grpdir == NULL) {
- if (errno == ENOENT)
- return 0;
- rc = -errno;
- VIR_ERROR(_("Unable to open %s (%d)"), grppath, errno);
- return rc;
- }
-
- for (;;) {
- char *path;
-
- errno = 0;
- ent = readdir(grpdir);
- if (ent == NULL) {
- if ((rc = -errno))
- VIR_ERROR(_("Failed to readdir for %s (%d)"), grppath, errno);
- break;
- }
-
- if (ent->d_name[0] == '.') continue;
- if (ent->d_type != DT_DIR) continue;
-
- if (virAsprintf(&path, "%s/%s", grppath, ent->d_name) == -1) {
- rc = -ENOMEM;
- break;
- }
- rc = virCgroupRemoveRecursively(path);
- VIR_FREE(path);
- if (rc != 0)
- break;
- }
- closedir(grpdir);
-
- VIR_DEBUG("Removing cgroup %s", grppath);
- if (rmdir(grppath) != 0 && errno != ENOENT) {
- rc = -errno;
- VIR_ERROR(_("Unable to remove %s (%d)"), grppath, errno);
- }
-
- return rc;
-}
-#else
-static int virCgroupRemoveRecursively(char *grppath ATTRIBUTE_UNUSED)
-{
- /* Claim no support */
- return -ENXIO;
-}
-#endif
-
-/**
- * virCgroupRemove:
- *
- * @group: The group to be removed
- *
- * It first removes all child groups recursively
- * in depth first order and then removes @group
- * because the presence of the child groups
- * prevents removing @group.
- *
- * Returns: 0 on success
- */
-int virCgroupRemove(virCgroupPtr group)
-{
- int rc = 0;
- int i;
- char *grppath = NULL;
-
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- /* Skip over controllers not mounted */
- if (!group->controllers[i].mountPoint)
- continue;
-
- if (virCgroupPathOfController(group,
- i,
- NULL,
- &grppath) != 0)
- continue;
-
- VIR_DEBUG("Removing cgroup %s and all child cgroups", grppath);
- rc = virCgroupRemoveRecursively(grppath);
- VIR_FREE(grppath);
- }
-
- return rc;
-}
-
-/**
- * virCgroupAddTask:
- *
- * @group: The cgroup to add a task to
- * @pid: The pid of the task to add
- *
- * Returns: 0 on success
- */
-int virCgroupAddTask(virCgroupPtr group, pid_t pid)
-{
- int rc = 0;
- int i;
-
- for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
- /* Skip over controllers not mounted */
- if (!group->controllers[i].mountPoint)
- continue;
-
- rc = virCgroupSetValueU64(group, i, "tasks", (unsigned long long)pid);
- if (rc != 0)
- break;
- }
-
- return rc;
-}
-
-/**
- * virCgroupAddTaskController:
- *
- * @group: The cgroup to add a task to
- * @pid: The pid of the task to add
- * @controller: The cgroup controller to be operated on
- *
- * Returns: 0 on success or -errno on failure
- */
-int virCgroupAddTaskController(virCgroupPtr group, pid_t pid, int controller)
-{
- if (controller < 0 || controller >= VIR_CGROUP_CONTROLLER_LAST)
- return -EINVAL;
-
- if (!group->controllers[controller].mountPoint)
- return -EINVAL;
-
- return virCgroupSetValueU64(group, controller, "tasks",
- (unsigned long long)pid);
-}
-
-
-static int virCgroupAddTaskStrController(virCgroupPtr group,
- const char *pidstr,
- int controller)
-{
- char *str = NULL, *cur = NULL, *next = NULL;
- unsigned long long p = 0;
- int rc = 0;
- char *endp;
-
- if (virAsprintf(&str, "%s", pidstr) < 0)
- return -1;
-
- cur = str;
- while (*cur != '\0') {
- rc = virStrToLong_ull(cur, &endp, 10, &p);
- if (rc != 0)
- goto cleanup;
-
- rc = virCgroupAddTaskController(group, p, controller);
- if (rc != 0)
- goto cleanup;
-
- next = strchr(cur, '\n');
- if (next) {
- cur = next + 1;
- *next = '\0';
- } else {
- break;
- }
- }
-
-cleanup:
- VIR_FREE(str);
- return rc;
-}
-
-/**
- * virCgroupMoveTask:
- *
- * @src_group: The source cgroup where all tasks are removed from
- * @dest_group: The destination where all tasks are added to
- * @controller: The cgroup controller to be operated on
- *
- * Returns: 0 on success or -errno on failure
- */
-int virCgroupMoveTask(virCgroupPtr src_group, virCgroupPtr dest_group,
- int controller)
-{
- int rc = 0, err = 0;
- char *content = NULL;
-
- if (controller < VIR_CGROUP_CONTROLLER_CPU ||
- controller > VIR_CGROUP_CONTROLLER_BLKIO)
- return -EINVAL;
-
- if (!src_group->controllers[controller].mountPoint ||
- !dest_group->controllers[controller].mountPoint) {
- VIR_WARN("no vm cgroup in controller %d", controller);
- return 0;
- }
-
- rc = virCgroupGetValueStr(src_group, controller, "tasks", &content);
- if (rc != 0)
- return rc;
-
- rc = virCgroupAddTaskStrController(dest_group, content, controller);
- if (rc != 0)
- goto cleanup;
-
- VIR_FREE(content);
-
- return 0;
-
-cleanup:
- /*
- * We don't need to recover dest_cgroup because cgroup will make sure
- * that one task only resides in one cgroup of the same controller.
- */
- err = virCgroupAddTaskStrController(src_group, content, controller);
- if (err != 0)
- VIR_ERROR(_("Cannot recover cgroup %s from %s"),
- src_group->controllers[controller].mountPoint,
- dest_group->controllers[controller].mountPoint);
- VIR_FREE(content);
-
- return rc;
-}
-
-/**
- * virCgroupForDriver:
- *
- * @name: name of this driver (e.g., xen, qemu, lxc)
- * @group: Pointer to returned virCgroupPtr
- *
- * Returns 0 on success
- */
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-int virCgroupForDriver(const char *name,
- virCgroupPtr *group,
- bool privileged,
- bool create)
-{
- int rc;
- char *path = NULL;
- virCgroupPtr rootgrp = NULL;
-
- rc = virCgroupAppRoot(privileged, &rootgrp, create);
- if (rc != 0)
- goto out;
-
- if (virAsprintf(&path, "%s/%s", rootgrp->path, name) < 0) {
- rc = -ENOMEM;
- goto out;
- }
-
- rc = virCgroupNew(path, group);
- VIR_FREE(path);
-
- if (rc == 0) {
- rc = virCgroupMakeGroup(rootgrp, *group, create, VIR_CGROUP_NONE);
- if (rc != 0)
- virCgroupFree(group);
- }
-
-out:
- virCgroupFree(&rootgrp);
-
- return rc;
-}
-#else
-int virCgroupForDriver(const char *name ATTRIBUTE_UNUSED,
- virCgroupPtr *group ATTRIBUTE_UNUSED,
- bool privileged ATTRIBUTE_UNUSED,
- bool create ATTRIBUTE_UNUSED)
-{
- /* Claim no support */
- return -ENXIO;
-}
-#endif
-
-/**
-* virCgroupGetAppRoot:
-*
-* @group: Pointer to returned virCgroupPtr
-*
-* Returns 0 on success
-*/
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-int virCgroupGetAppRoot(virCgroupPtr *group)
-{
- return virCgroupNew("/", group);
-}
-#else
-int virCgroupGetAppRoot(virCgroupPtr *group ATTRIBUTE_UNUSED)
-{
- return -ENXIO;
-}
-#endif
-
-/**
- * virCgroupForDomain:
- *
- * @driver: group for driver owning the domain
- * @name: name of the domain
- * @group: Pointer to returned virCgroupPtr
- *
- * Returns 0 on success
- */
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-int virCgroupForDomain(virCgroupPtr driver,
- const char *name,
- virCgroupPtr *group,
- bool create)
-{
- int rc;
- char *path;
-
- if (driver == NULL)
- return -EINVAL;
-
- if (virAsprintf(&path, "%s/%s", driver->path, name) < 0)
- return -ENOMEM;
-
- rc = virCgroupNew(path, group);
- VIR_FREE(path);
-
- if (rc == 0) {
- /*
- * Create a cgroup with memory.use_hierarchy enabled to
- * surely account memory usage of lxc with ns subsystem
- * enabled. (To be exact, memory and ns subsystems are
- * enabled at the same time.)
- *
- * The reason why doing it here, not a upper group, say
- * a group for driver, is to avoid overhead to track
- * cumulative usage that we don't need.
- */
- rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_MEM_HIERACHY);
- if (rc != 0)
- virCgroupFree(group);
- }
-
- return rc;
-}
-#else
-int virCgroupForDomain(virCgroupPtr driver ATTRIBUTE_UNUSED,
- const char *name ATTRIBUTE_UNUSED,
- virCgroupPtr *group ATTRIBUTE_UNUSED,
- bool create ATTRIBUTE_UNUSED)
-{
- return -ENXIO;
-}
-#endif
-
-/**
- * virCgroupForVcpu:
- *
- * @driver: group for the domain
- * @vcpuid: id of the vcpu
- * @group: Pointer to returned virCgroupPtr
- *
- * Returns 0 on success
- */
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-int virCgroupForVcpu(virCgroupPtr driver,
- int vcpuid,
- virCgroupPtr *group,
- bool create)
-{
- int rc;
- char *path;
-
- if (driver == NULL)
- return -EINVAL;
-
- if (virAsprintf(&path, "%s/vcpu%d", driver->path, vcpuid) < 0)
- return -ENOMEM;
-
- rc = virCgroupNew(path, group);
- VIR_FREE(path);
-
- if (rc == 0) {
- rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_VCPU);
- if (rc != 0)
- virCgroupFree(group);
- }
-
- return rc;
-}
-#else
-int virCgroupForVcpu(virCgroupPtr driver ATTRIBUTE_UNUSED,
- int vcpuid ATTRIBUTE_UNUSED,
- virCgroupPtr *group ATTRIBUTE_UNUSED,
- bool create ATTRIBUTE_UNUSED)
-{
- return -ENXIO;
-}
-#endif
-
-/**
- * virCgroupForEmulator:
- *
- * @driver: group for the domain
- * @group: Pointer to returned virCgroupPtr
- *
- * Returns: 0 on success or -errno on failure
- */
-#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-int virCgroupForEmulator(virCgroupPtr driver,
- virCgroupPtr *group,
- bool create)
-{
- int rc;
- char *path;
-
- if (driver == NULL)
- return -EINVAL;
-
- if (virAsprintf(&path, "%s/emulator", driver->path) < 0)
- return -ENOMEM;
-
- rc = virCgroupNew(path, group);
- VIR_FREE(path);
-
- if (rc == 0) {
- rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_VCPU);
- if (rc != 0)
- virCgroupFree(group);
- }
-
- return rc;
-}
-#else
-int virCgroupForEmulator(virCgroupPtr driver ATTRIBUTE_UNUSED,
- virCgroupPtr *group ATTRIBUTE_UNUSED,
- bool create ATTRIBUTE_UNUSED)
-{
- return -ENXIO;
-}
-
-#endif
-/**
- * virCgroupSetBlkioWeight:
- *
- * @group: The cgroup to change io weight for
- * @weight: The Weight for this cgroup
- *
- * Returns: 0 on success
- */
-int virCgroupSetBlkioWeight(virCgroupPtr group, unsigned int weight)
-{
- if (weight > 1000 || weight < 100)
- return -EINVAL;
-
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_BLKIO,
- "blkio.weight",
- weight);
-}
-
-/**
- * virCgroupGetBlkioWeight:
- *
- * @group: The cgroup to get weight for
- * @Weight: Pointer to returned weight
- *
- * Returns: 0 on success
- */
-int virCgroupGetBlkioWeight(virCgroupPtr group, unsigned int *weight)
-{
- unsigned long long tmp;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_BLKIO,
- "blkio.weight", &tmp);
- if (ret == 0)
- *weight = tmp;
- return ret;
-}
-
-/**
- * virCgroupSetBlkioDeviceWeight:
- *
- * @group: The cgroup to change io device weight device for
- * @path: The device with a weight to alter
- * @weight: The new device weight (100-1000), or 0 to clear
- *
- * device_weight is treated as a write-only parameter, so
- * there isn't a getter counterpart.
- *
- * Returns: 0 on success, -errno on failure
- */
-#if defined(major) && defined(minor)
-int virCgroupSetBlkioDeviceWeight(virCgroupPtr group,
- const char *path,
- unsigned int weight)
-{
- char *str;
- struct stat sb;
- int ret;
-
- if (weight && (weight > 1000 || weight < 100))
- return -EINVAL;
-
- if (stat(path, &sb) < 0)
- return -errno;
-
- if (!S_ISBLK(sb.st_mode))
- return -EINVAL;
-
- if (virAsprintf(&str, "%d:%d %d", major(sb.st_rdev), minor(sb.st_rdev),
- weight) < 0)
- return -errno;
-
- ret = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_BLKIO,
- "blkio.weight_device",
- str);
- VIR_FREE(str);
- return ret;
-}
-#else
-int
-virCgroupSetBlkioDeviceWeight(virCgroupPtr group ATTRIBUTE_UNUSED,
- const char *path ATTRIBUTE_UNUSED,
- unsigned int weight ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-#endif
-
-/**
- * virCgroupSetMemory:
- *
- * @group: The cgroup to change memory for
- * @kb: The memory amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupSetMemory(virCgroupPtr group, unsigned long long kb)
-{
- unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
-
- if (kb > maxkb)
- return -EINVAL;
- else if (kb == maxkb)
- return virCgroupSetValueI64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.limit_in_bytes",
- -1);
- else
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.limit_in_bytes",
- kb << 10);
-}
-
-/**
- * virCgroupGetMemoryUsage:
- *
- * @group: The cgroup to change memory for
- * @kb: Pointer to returned used memory in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupGetMemoryUsage(virCgroupPtr group, unsigned long *kb)
-{
- long long unsigned int usage_in_bytes;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.usage_in_bytes", &usage_in_bytes);
- if (ret == 0)
- *kb = (unsigned long) usage_in_bytes >> 10;
- return ret;
-}
-
-/**
- * virCgroupSetMemoryHardLimit:
- *
- * @group: The cgroup to change memory hard limit for
- * @kb: The memory amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupSetMemoryHardLimit(virCgroupPtr group, unsigned long long kb)
-{
- return virCgroupSetMemory(group, kb);
-}
-
-/**
- * virCgroupGetMemoryHardLimit:
- *
- * @group: The cgroup to get the memory hard limit for
- * @kb: The memory amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupGetMemoryHardLimit(virCgroupPtr group, unsigned long long *kb)
-{
- long long unsigned int limit_in_bytes;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.limit_in_bytes", &limit_in_bytes);
- if (ret == 0)
- *kb = limit_in_bytes >> 10;
- return ret;
-}
-
-/**
- * virCgroupSetMemorySoftLimit:
- *
- * @group: The cgroup to change memory soft limit for
- * @kb: The memory amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupSetMemorySoftLimit(virCgroupPtr group, unsigned long long kb)
-{
- unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
-
- if (kb > maxkb)
- return -EINVAL;
- else if (kb == maxkb)
- return virCgroupSetValueI64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.soft_limit_in_bytes",
- -1);
- else
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.soft_limit_in_bytes",
- kb << 10);
-}
-
-
-/**
- * virCgroupGetMemorySoftLimit:
- *
- * @group: The cgroup to get the memory soft limit for
- * @kb: The memory amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupGetMemorySoftLimit(virCgroupPtr group, unsigned long long *kb)
-{
- long long unsigned int limit_in_bytes;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.soft_limit_in_bytes", &limit_in_bytes);
- if (ret == 0)
- *kb = limit_in_bytes >> 10;
- return ret;
-}
-
-/**
- * virCgroupSetMemSwapHardLimit:
- *
- * @group: The cgroup to change mem+swap hard limit for
- * @kb: The mem+swap amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupSetMemSwapHardLimit(virCgroupPtr group, unsigned long long kb)
-{
- unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
-
- if (kb > maxkb)
- return -EINVAL;
- else if (kb == maxkb)
- return virCgroupSetValueI64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.memsw.limit_in_bytes",
- -1);
- else
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.memsw.limit_in_bytes",
- kb << 10);
-}
-
-/**
- * virCgroupGetMemSwapHardLimit:
- *
- * @group: The cgroup to get mem+swap hard limit for
- * @kb: The mem+swap amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupGetMemSwapHardLimit(virCgroupPtr group, unsigned long long *kb)
-{
- long long unsigned int limit_in_bytes;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.memsw.limit_in_bytes", &limit_in_bytes);
- if (ret == 0)
- *kb = limit_in_bytes >> 10;
- return ret;
-}
-
-/**
- * virCgroupGetMemSwapUsage:
- *
- * @group: The cgroup to get mem+swap usage for
- * @kb: The mem+swap amount in kilobytes
- *
- * Returns: 0 on success
- */
-int virCgroupGetMemSwapUsage(virCgroupPtr group, unsigned long long *kb)
-{
- long long unsigned int usage_in_bytes;
- int ret;
- ret = virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_MEMORY,
- "memory.memsw.usage_in_bytes", &usage_in_bytes);
- if (ret == 0)
- *kb = usage_in_bytes >> 10;
- return ret;
-}
-
-/**
- * virCgroupSetCpusetMems:
- *
- * @group: The cgroup to set cpuset.mems for
- * @mems: the numa nodes to set
- *
- * Returns: 0 on success
- */
-int virCgroupSetCpusetMems(virCgroupPtr group, const char *mems)
-{
- return virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_CPUSET,
- "cpuset.mems",
- mems);
-}
-
-/**
- * virCgroupGetCpusetMems:
- *
- * @group: The cgroup to get cpuset.mems for
- * @mems: the numa nodes to get
- *
- * Returns: 0 on success
- */
-int virCgroupGetCpusetMems(virCgroupPtr group, char **mems)
-{
- return virCgroupGetValueStr(group,
- VIR_CGROUP_CONTROLLER_CPUSET,
- "cpuset.mems",
- mems);
-}
-
-/**
- * virCgroupSetCpusetCpus:
- *
- * @group: The cgroup to set cpuset.cpus for
- * @cpus: the cpus to set
- *
- * Retuens: 0 on success
- */
-int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus)
-{
- return virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_CPUSET,
- "cpuset.cpus",
- cpus);
-}
-
-/**
- * virCgroupGetCpusetCpus:
- *
- * @group: The cgroup to get cpuset.cpus for
- * @cpus: the cpus to get
- *
- * Retuens: 0 on success
- */
-int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus)
-{
- return virCgroupGetValueStr(group,
- VIR_CGROUP_CONTROLLER_CPUSET,
- "cpuset.cpus",
- cpus);
-}
-
-/**
- * virCgroupDenyAllDevices:
- *
- * @group: The cgroup to deny all permissions, for all devices
- *
- * Returns: 0 on success
- */
-int virCgroupDenyAllDevices(virCgroupPtr group)
-{
- return virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_DEVICES,
- "devices.deny",
- "a");
-}
-
-/**
- * virCgroupAllowDevice:
- *
- * @group: The cgroup to allow a device for
- * @type: The device type (i.e., 'c' or 'b')
- * @major: The major number of the device
- * @minor: The minor number of the device
- * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
- *
- * Returns: 0 on success
- */
-int virCgroupAllowDevice(virCgroupPtr group, char type, int major, int minor,
- int perms)
-{
- int rc;
- char *devstr = NULL;
-
- if (virAsprintf(&devstr, "%c %i:%i %s%s%s", type, major, minor,
- perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
- perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
- perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
- rc = -ENOMEM;
- goto out;
- }
-
- rc = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_DEVICES,
- "devices.allow",
- devstr);
-out:
- VIR_FREE(devstr);
-
- return rc;
-}
-
-/**
- * virCgroupAllowDeviceMajor:
- *
- * @group: The cgroup to allow an entire device major type for
- * @type: The device type (i.e., 'c' or 'b')
- * @major: The major number of the device type
- * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
- *
- * Returns: 0 on success
- */
-int virCgroupAllowDeviceMajor(virCgroupPtr group, char type, int major,
- int perms)
-{
- int rc;
- char *devstr = NULL;
-
- if (virAsprintf(&devstr, "%c %i:* %s%s%s", type, major,
- perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
- perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
- perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
- rc = -ENOMEM;
- goto out;
- }
-
- rc = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_DEVICES,
- "devices.allow",
- devstr);
- out:
- VIR_FREE(devstr);
-
- return rc;
-}
-
-/**
- * virCgroupAllowDevicePath:
- *
- * @group: The cgroup to allow the device for
- * @path: the device to allow
- * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
- *
- * Queries the type of device and its major/minor number, and
- * adds that to the cgroup ACL
- *
- * Returns: 0 on success, 1 if path exists but is not a device, or
- * negative errno value on failure
- */
-#if defined(major) && defined(minor)
-int virCgroupAllowDevicePath(virCgroupPtr group, const char *path, int perms)
-{
- struct stat sb;
-
- if (stat(path, &sb) < 0)
- return -errno;
-
- if (!S_ISCHR(sb.st_mode) && !S_ISBLK(sb.st_mode))
- return 1;
-
- return virCgroupAllowDevice(group,
- S_ISCHR(sb.st_mode) ? 'c' : 'b',
- major(sb.st_rdev),
- minor(sb.st_rdev),
- perms);
-}
-#else
-int virCgroupAllowDevicePath(virCgroupPtr group ATTRIBUTE_UNUSED,
- const char *path ATTRIBUTE_UNUSED,
- int perms ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-#endif
-
-
-/**
- * virCgroupDenyDevice:
- *
- * @group: The cgroup to deny a device for
- * @type: The device type (i.e., 'c' or 'b')
- * @major: The major number of the device
- * @minor: The minor number of the device
- * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to deny
- *
- * Returns: 0 on success
- */
-int virCgroupDenyDevice(virCgroupPtr group, char type, int major, int minor,
- int perms)
-{
- int rc;
- char *devstr = NULL;
-
- if (virAsprintf(&devstr, "%c %i:%i %s%s%s", type, major, minor,
- perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
- perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
- perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
- rc = -ENOMEM;
- goto out;
- }
-
- rc = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_DEVICES,
- "devices.deny",
- devstr);
-out:
- VIR_FREE(devstr);
-
- return rc;
-}
-
-/**
- * virCgroupDenyDeviceMajor:
- *
- * @group: The cgroup to deny an entire device major type for
- * @type: The device type (i.e., 'c' or 'b')
- * @major: The major number of the device type
- * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to deny
- *
- * Returns: 0 on success
- */
-int virCgroupDenyDeviceMajor(virCgroupPtr group, char type, int major,
- int perms)
-{
- int rc;
- char *devstr = NULL;
-
- if (virAsprintf(&devstr, "%c %i:* %s%s%s", type, major,
- perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
- perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
- perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
- rc = -ENOMEM;
- goto out;
- }
-
- rc = virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_DEVICES,
- "devices.deny",
- devstr);
- out:
- VIR_FREE(devstr);
-
- return rc;
-}
-
-#if defined(major) && defined(minor)
-int virCgroupDenyDevicePath(virCgroupPtr group, const char *path, int perms)
-{
- struct stat sb;
-
- if (stat(path, &sb) < 0)
- return -errno;
-
- if (!S_ISCHR(sb.st_mode) && !S_ISBLK(sb.st_mode))
- return 1;
-
- return virCgroupDenyDevice(group,
- S_ISCHR(sb.st_mode) ? 'c' : 'b',
- major(sb.st_rdev),
- minor(sb.st_rdev),
- perms);
-}
-#else
-int virCgroupDenyDevicePath(virCgroupPtr group ATTRIBUTE_UNUSED,
- const char *path ATTRIBUTE_UNUSED,
- int perms ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-#endif
-
-int virCgroupSetCpuShares(virCgroupPtr group, unsigned long long shares)
-{
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.shares", shares);
-}
-
-int virCgroupGetCpuShares(virCgroupPtr group, unsigned long long *shares)
-{
- return virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.shares", shares);
-}
-
-/**
- * virCgroupSetCpuCfsPeriod:
- *
- * @group: The cgroup to change cpu.cfs_period_us for
- * @cfs_period: The bandwidth period in usecs
- *
- * Returns: 0 on success
- */
-int virCgroupSetCpuCfsPeriod(virCgroupPtr group, unsigned long long cfs_period)
-{
- /* The cfs_period shoule be greater or equal than 1ms, and less or equal
- * than 1s.
- */
- if (cfs_period < 1000 || cfs_period > 1000000)
- return -EINVAL;
-
- return virCgroupSetValueU64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.cfs_period_us", cfs_period);
-}
-
-/**
- * virCgroupGetCpuCfsPeriod:
- *
- * @group: The cgroup to get cpu.cfs_period_us for
- * @cfs_period: Pointer to the returned bandwidth period in usecs
- *
- * Returns: 0 on success
- */
-int virCgroupGetCpuCfsPeriod(virCgroupPtr group, unsigned long long *cfs_period)
-{
- return virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.cfs_period_us", cfs_period);
-}
-
-/**
- * virCgroupSetCpuCfsQuota:
- *
- * @group: The cgroup to change cpu.cfs_quota_us for
- * @cfs_quota: the cpu bandwidth (in usecs) that this tg will be allowed to
- * consume over period
- *
- * Returns: 0 on success
- */
-int virCgroupSetCpuCfsQuota(virCgroupPtr group, long long cfs_quota)
-{
- if (cfs_quota >= 0) {
- /* The cfs_quota shoule be greater or equal than 1ms */
- if (cfs_quota < 1000)
- return -EINVAL;
-
- /* check overflow */
- if (cfs_quota > ULLONG_MAX / 1000)
- return -EINVAL;
- }
-
- return virCgroupSetValueI64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.cfs_quota_us", cfs_quota);
-}
-
-/**
- * virCgroupGetCpuCfsQuota:
- *
- * @group: The cgroup to get cpu.cfs_quota_us for
- * @cfs_quota: Pointer to the returned cpu bandwidth (in usecs) that this tg
- * will be allowed to consume over period
- *
- * Returns: 0 on success
- */
-int virCgroupGetCpuCfsQuota(virCgroupPtr group, long long *cfs_quota)
-{
- return virCgroupGetValueI64(group,
- VIR_CGROUP_CONTROLLER_CPU,
- "cpu.cfs_quota_us", cfs_quota);
-}
-
-int virCgroupGetCpuacctUsage(virCgroupPtr group, unsigned long long *usage)
-{
- return virCgroupGetValueU64(group,
- VIR_CGROUP_CONTROLLER_CPUACCT,
- "cpuacct.usage", usage);
-}
-
-int virCgroupGetCpuacctPercpuUsage(virCgroupPtr group, char **usage)
-{
- return virCgroupGetValueStr(group, VIR_CGROUP_CONTROLLER_CPUACCT,
- "cpuacct.usage_percpu", usage);
-}
-
-#ifdef _SC_CLK_TCK
-int virCgroupGetCpuacctStat(virCgroupPtr group, unsigned long long *user,
- unsigned long long *sys)
-{
- char *str;
- char *p;
- int ret;
- static double scale = -1.0;
-
- if ((ret = virCgroupGetValueStr(group, VIR_CGROUP_CONTROLLER_CPUACCT,
- "cpuacct.stat", &str)) < 0)
- return ret;
- if (!(p = STRSKIP(str, "user ")) ||
- virStrToLong_ull(p, &p, 10, user) < 0 ||
- !(p = STRSKIP(p, "\nsystem ")) ||
- virStrToLong_ull(p, NULL, 10, sys) < 0) {
- ret = -EINVAL;
- goto cleanup;
- }
- /* times reported are in system ticks (generally 100 Hz), but that
- * rate can theoretically vary between machines. Scale things
- * into approximate nanoseconds. */
- if (scale < 0) {
- long ticks_per_sec = sysconf(_SC_CLK_TCK);
- if (ticks_per_sec == -1) {
- ret = -errno;
- goto cleanup;
- }
- scale = 1000000000.0 / ticks_per_sec;
- }
- *user *= scale;
- *sys *= scale;
-
- ret = 0;
-cleanup:
- VIR_FREE(str);
- return ret;
-}
-#else
-int virCgroupGetCpuacctStat(virCgroupPtr group ATTRIBUTE_UNUSED,
- unsigned long long *user ATTRIBUTE_UNUSED,
- unsigned long long *sys ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-#endif
-
-int virCgroupSetFreezerState(virCgroupPtr group, const char *state)
-{
- return virCgroupSetValueStr(group,
- VIR_CGROUP_CONTROLLER_FREEZER,
- "freezer.state", state);
-}
-
-int virCgroupGetFreezerState(virCgroupPtr group, char **state)
-{
- return virCgroupGetValueStr(group,
- VIR_CGROUP_CONTROLLER_FREEZER,
- "freezer.state", state);
-}
-
-
-#if defined HAVE_KILL && defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
-static int virCgroupKillInternal(virCgroupPtr group, int signum, virHashTablePtr pids)
-{
- int rc;
- int killedAny = 0;
- char *keypath = NULL;
- bool done = false;
- FILE *fp = NULL;
- VIR_DEBUG("group=%p path=%s signum=%d pids=%p",
- group, group->path, signum, pids);
-
- rc = virCgroupPathOfController(group, -1, "tasks", &keypath);
- if (rc != 0) {
- VIR_DEBUG("No path of %s, tasks", group->path);
- return rc;
- }
-
- /* PIDs may be forking as we kill them, so loop
- * until there are no new PIDs found
- */
- while (!done) {
- done = true;
- if (!(fp = fopen(keypath, "r"))) {
- rc = -errno;
- VIR_DEBUG("Failed to read %s: %m\n", keypath);
- goto cleanup;
- } else {
- while (!feof(fp)) {
- unsigned long pid_value;
- if (fscanf(fp, "%lu", &pid_value) != 1) {
- if (feof(fp))
- break;
- rc = -errno;
- VIR_DEBUG("Failed to read %s: %m\n", keypath);
- goto cleanup;
- }
- if (virHashLookup(pids, (void*)pid_value))
- continue;
-
- VIR_DEBUG("pid=%lu", pid_value);
- /* Cgroups is a Linux concept, so this cast is safe. */
- if (kill((pid_t)pid_value, signum) < 0) {
- if (errno != ESRCH) {
- rc = -errno;
- goto cleanup;
- }
- /* Leave RC == 0 since we didn't kill one */
- } else {
- killedAny = 1;
- done = false;
- }
-
- ignore_value(virHashAddEntry(pids, (void*)pid_value, (void*)1));
- }
- VIR_FORCE_FCLOSE(fp);
- }
- }
-
- rc = killedAny ? 1 : 0;
-
-cleanup:
- VIR_FREE(keypath);
- VIR_FORCE_FCLOSE(fp);
-
- return rc;
-}
-
-
-static uint32_t virCgroupPidCode(const void *name, uint32_t seed)
-{
- unsigned long pid_value = (unsigned long)(intptr_t)name;
- return virHashCodeGen(&pid_value, sizeof(pid_value), seed);
-}
-static bool virCgroupPidEqual(const void *namea, const void *nameb)
-{
- return namea == nameb;
-}
-static void *virCgroupPidCopy(const void *name)
-{
- return (void*)name;
-}
-
-/*
- * Returns
- * < 0 : errno that occurred
- * 0 : no PIDs killed
- * 1 : at least one PID killed
- */
-int virCgroupKill(virCgroupPtr group, int signum)
-{
- VIR_DEBUG("group=%p path=%s signum=%d", group, group->path, signum);
- int rc;
- /* The 'tasks' file in cgroups can contain duplicated
- * pids, so we use a hash to track which we've already
- * killed.
- */
- virHashTablePtr pids = virHashCreateFull(100,
- NULL,
- virCgroupPidCode,
- virCgroupPidEqual,
- virCgroupPidCopy,
- NULL);
-
- rc = virCgroupKillInternal(group, signum, pids);
-
- virHashFree(pids);
-
- return rc;
-}
-
-
-static int virCgroupKillRecursiveInternal(virCgroupPtr group, int signum, virHashTablePtr pids, bool dormdir)
-{
- int rc;
- int killedAny = 0;
- char *keypath = NULL;
- DIR *dp;
- virCgroupPtr subgroup = NULL;
- struct dirent *ent;
- VIR_DEBUG("group=%p path=%s signum=%d pids=%p", group, group->path, signum, pids);
-
- rc = virCgroupPathOfController(group, -1, "", &keypath);
- if (rc != 0) {
- VIR_DEBUG("No path of %s, tasks", group->path);
- return rc;
- }
-
- if ((rc = virCgroupKillInternal(group, signum, pids)) != 0)
- return rc;
-
- VIR_DEBUG("Iterate over children of %s", keypath);
- if (!(dp = opendir(keypath))) {
- rc = -errno;
- return rc;
- }
-
- while ((ent = readdir(dp))) {
- char *subpath;
-
- if (STREQ(ent->d_name, "."))
- continue;
- if (STREQ(ent->d_name, ".."))
- continue;
- if (ent->d_type != DT_DIR)
- continue;
-
- VIR_DEBUG("Process subdir %s", ent->d_name);
- if (virAsprintf(&subpath, "%s/%s", group->path, ent->d_name) < 0) {
- rc = -ENOMEM;
- goto cleanup;
- }
-
- if ((rc = virCgroupNew(subpath, &subgroup)) != 0)
- goto cleanup;
-
- if ((rc = virCgroupKillRecursiveInternal(subgroup, signum, pids, true)) < 0)
- goto cleanup;
- if (rc == 1)
- killedAny = 1;
-
- if (dormdir)
- virCgroupRemove(subgroup);
-
- virCgroupFree(&subgroup);
- }
-
- rc = killedAny;
-
-cleanup:
- virCgroupFree(&subgroup);
- closedir(dp);
-
- return rc;
-}
-
-int virCgroupKillRecursive(virCgroupPtr group, int signum)
-{
- int rc;
- VIR_DEBUG("group=%p path=%s signum=%d", group, group->path, signum);
- virHashTablePtr pids = virHashCreateFull(100,
- NULL,
- virCgroupPidCode,
- virCgroupPidEqual,
- virCgroupPidCopy,
- NULL);
-
- rc = virCgroupKillRecursiveInternal(group, signum, pids, false);
-
- virHashFree(pids);
-
- return rc;
-}
-
-
-int virCgroupKillPainfully(virCgroupPtr group)
-{
- int i;
- int rc;
- VIR_DEBUG("cgroup=%p path=%s", group, group->path);
- for (i = 0 ; i < 15 ; i++) {
- int signum;
- if (i == 0)
- signum = SIGTERM;
- else if (i == 8)
- signum = SIGKILL;
- else
- signum = 0; /* Just check for existence */
-
- rc = virCgroupKillRecursive(group, signum);
- VIR_DEBUG("Iteration %d rc=%d", i, rc);
- /* If rc == -1 we hit error, if 0 we ran out of PIDs */
- if (rc <= 0)
- break;
-
- usleep(200 * 1000);
- }
- VIR_DEBUG("Complete %d", rc);
- return rc;
-}
-
-#else /* !(HAVE_KILL, HAVE_MNTENT_H, HAVE_GETMNTENT_R) */
-int virCgroupKill(virCgroupPtr group ATTRIBUTE_UNUSED,
- int signum ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-int virCgroupKillRecursive(virCgroupPtr group ATTRIBUTE_UNUSED,
- int signum ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-
-int virCgroupKillPainfully(virCgroupPtr group ATTRIBUTE_UNUSED)
-{
- return -ENOSYS;
-}
-#endif /* HAVE_KILL, HAVE_MNTENT_H, HAVE_GETMNTENT_R */
diff --git a/src/util/cgroup.h b/src/util/cgroup.h
deleted file mode 100644
index fc9e409..0000000
--- a/src/util/cgroup.h
+++ /dev/null
@@ -1,167 +0,0 @@
-/*
- * cgroup.h: Interface to tools for managing cgroups
- *
- * Copyright (C) 2011-2012 Red Hat, Inc.
- * Copyright IBM Corp. 2008
- *
- * 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/>.
- *
- * Authors:
- * Dan Smith <danms@xxxxxxxxxx>
- */
-
-#ifndef CGROUP_H
-# define CGROUP_H
-
-struct virCgroup;
-typedef struct virCgroup *virCgroupPtr;
-
-enum {
- VIR_CGROUP_CONTROLLER_CPU,
- VIR_CGROUP_CONTROLLER_CPUACCT,
- VIR_CGROUP_CONTROLLER_CPUSET,
- VIR_CGROUP_CONTROLLER_MEMORY,
- VIR_CGROUP_CONTROLLER_DEVICES,
- VIR_CGROUP_CONTROLLER_FREEZER,
- VIR_CGROUP_CONTROLLER_BLKIO,
-
- VIR_CGROUP_CONTROLLER_LAST
-};
-
-VIR_ENUM_DECL(virCgroupController);
-
-int virCgroupForDriver(const char *name,
- virCgroupPtr *group,
- bool privileged,
- bool create);
-
-int virCgroupGetAppRoot(virCgroupPtr *group);
-
-int virCgroupForDomain(virCgroupPtr driver,
- const char *name,
- virCgroupPtr *group,
- bool create);
-
-int virCgroupForVcpu(virCgroupPtr driver,
- int vcpuid,
- virCgroupPtr *group,
- bool create);
-
-int virCgroupForEmulator(virCgroupPtr driver,
- virCgroupPtr *group,
- bool create);
-
-int virCgroupPathOfController(virCgroupPtr group,
- int controller,
- const char *key,
- char **path);
-
-int virCgroupAddTask(virCgroupPtr group, pid_t pid);
-
-int virCgroupAddTaskController(virCgroupPtr group,
- pid_t pid,
- int controller);
-
-int virCgroupMoveTask(virCgroupPtr src_group,
- virCgroupPtr dest_group,
- int controller);
-
-int virCgroupSetBlkioWeight(virCgroupPtr group, unsigned int weight);
-int virCgroupGetBlkioWeight(virCgroupPtr group, unsigned int *weight);
-
-int virCgroupSetBlkioDeviceWeight(virCgroupPtr group,
- const char *path,
- unsigned int weight);
-
-int virCgroupSetMemory(virCgroupPtr group, unsigned long long kb);
-int virCgroupGetMemoryUsage(virCgroupPtr group, unsigned long *kb);
-
-int virCgroupSetMemoryHardLimit(virCgroupPtr group, unsigned long long kb);
-int virCgroupGetMemoryHardLimit(virCgroupPtr group, unsigned long long *kb);
-int virCgroupSetMemorySoftLimit(virCgroupPtr group, unsigned long long kb);
-int virCgroupGetMemorySoftLimit(virCgroupPtr group, unsigned long long *kb);
-int virCgroupSetMemSwapHardLimit(virCgroupPtr group, unsigned long long kb);
-int virCgroupGetMemSwapHardLimit(virCgroupPtr group, unsigned long long *kb);
-int virCgroupGetMemSwapUsage(virCgroupPtr group, unsigned long long *kb);
-
-enum {
- VIR_CGROUP_DEVICE_READ = 1,
- VIR_CGROUP_DEVICE_WRITE = 2,
- VIR_CGROUP_DEVICE_MKNOD = 4,
- VIR_CGROUP_DEVICE_RW = VIR_CGROUP_DEVICE_READ | VIR_CGROUP_DEVICE_WRITE,
- VIR_CGROUP_DEVICE_RWM = VIR_CGROUP_DEVICE_RW | VIR_CGROUP_DEVICE_MKNOD,
-};
-
-int virCgroupDenyAllDevices(virCgroupPtr group);
-
-int virCgroupAllowDevice(virCgroupPtr group,
- char type,
- int major,
- int minor,
- int perms);
-int virCgroupAllowDeviceMajor(virCgroupPtr group,
- char type,
- int major,
- int perms);
-int virCgroupAllowDevicePath(virCgroupPtr group,
- const char *path,
- int perms);
-
-int virCgroupDenyDevice(virCgroupPtr group,
- char type,
- int major,
- int minor,
- int perms);
-int virCgroupDenyDeviceMajor(virCgroupPtr group,
- char type,
- int major,
- int perms);
-int virCgroupDenyDevicePath(virCgroupPtr group,
- const char *path,
- int perms);
-
-int virCgroupSetCpuShares(virCgroupPtr group, unsigned long long shares);
-int virCgroupGetCpuShares(virCgroupPtr group, unsigned long long *shares);
-
-int virCgroupSetCpuCfsPeriod(virCgroupPtr group, unsigned long long cfs_period);
-int virCgroupGetCpuCfsPeriod(virCgroupPtr group, unsigned long long *cfs_period);
-
-int virCgroupSetCpuCfsQuota(virCgroupPtr group, long long cfs_quota);
-int virCgroupGetCpuCfsQuota(virCgroupPtr group, long long *cfs_quota);
-
-int virCgroupGetCpuacctUsage(virCgroupPtr group, unsigned long long *usage);
-int virCgroupGetCpuacctPercpuUsage(virCgroupPtr group, char **usage);
-int virCgroupGetCpuacctStat(virCgroupPtr group, unsigned long long *user,
- unsigned long long *sys);
-
-int virCgroupSetFreezerState(virCgroupPtr group, const char *state);
-int virCgroupGetFreezerState(virCgroupPtr group, char **state);
-
-int virCgroupSetCpusetMems(virCgroupPtr group, const char *mems);
-int virCgroupGetCpusetMems(virCgroupPtr group, char **mems);
-
-int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus);
-int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus);
-
-int virCgroupRemove(virCgroupPtr group);
-
-void virCgroupFree(virCgroupPtr *group);
-bool virCgroupMounted(virCgroupPtr cgroup, int controller);
-
-int virCgroupKill(virCgroupPtr group, int signum);
-int virCgroupKillRecursive(virCgroupPtr group, int signum);
-int virCgroupKillPainfully(virCgroupPtr group);
-
-#endif /* CGROUP_H */
diff --git a/src/util/vircgroup.c b/src/util/vircgroup.c
new file mode 100644
index 0000000..e9fc67f
--- /dev/null
+++ b/src/util/vircgroup.c
@@ -0,0 +1,2099 @@
+/*
+ * vircgroup.c: methods for managing control cgroups
+ *
+ * Copyright (C) 2010-2012 Red Hat, Inc.
+ * Copyright IBM Corp. 2008
+ *
+ * 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/>.
+ *
+ * Authors:
+ * Dan Smith <danms@xxxxxxxxxx>
+ */
+#include <config.h>
+
+#include <stdio.h>
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+# include <mntent.h>
+#endif
+#include <fcntl.h>
+#include <string.h>
+#include <errno.h>
+#include <stdlib.h>
+#include <sys/stat.h>
+#include <sys/types.h>
+#include <signal.h>
+#include <libgen.h>
+#include <dirent.h>
+
+#include "internal.h"
+#include "util.h"
+#include "memory.h"
+#include "vircgroup.h"
+#include "logging.h"
+#include "virfile.h"
+#include "virhash.h"
+#include "virhashcode.h"
+
+#define CGROUP_MAX_VAL 512
+
+VIR_ENUM_IMPL(virCgroupController, VIR_CGROUP_CONTROLLER_LAST,
+ "cpu", "cpuacct", "cpuset", "memory", "devices",
+ "freezer", "blkio");
+
+struct virCgroupController {
+ int type;
+ char *mountPoint;
+ char *placement;
+};
+
+struct virCgroup {
+ char *path;
+
+ struct virCgroupController controllers[VIR_CGROUP_CONTROLLER_LAST];
+};
+
+typedef enum {
+ VIR_CGROUP_NONE = 0, /* create subdir under each cgroup if possible. */
+ VIR_CGROUP_MEM_HIERACHY = 1 << 0, /* call virCgroupSetMemoryUseHierarchy
+ * before creating subcgroups and
+ * attaching tasks
+ */
+ VIR_CGROUP_VCPU = 1 << 1, /* create subdir only under the cgroup cpu,
+ * cpuacct and cpuset if possible. */
+} virCgroupFlags;
+
+/**
+ * virCgroupFree:
+ *
+ * @group: The group structure to free
+ */
+void virCgroupFree(virCgroupPtr *group)
+{
+ int i;
+
+ if (*group == NULL)
+ return;
+
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ VIR_FREE((*group)->controllers[i].mountPoint);
+ VIR_FREE((*group)->controllers[i].placement);
+ }
+
+ VIR_FREE((*group)->path);
+ VIR_FREE(*group);
+}
+
+/**
+ * virCgroupMounted: query whether a cgroup subsystem is mounted or not
+ *
+ * @cgroup: The group structure to be queried
+ * @controller: cgroup subsystem id
+ *
+ * Returns true if a cgroup is subsystem is mounted.
+ */
+bool virCgroupMounted(virCgroupPtr cgroup, int controller)
+{
+ return cgroup->controllers[controller].mountPoint != NULL;
+}
+
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+/*
+ * Process /proc/mounts figuring out what controllers are
+ * mounted and where
+ */
+static int virCgroupDetectMounts(virCgroupPtr group)
+{
+ int i;
+ FILE *mounts = NULL;
+ struct mntent entry;
+ char buf[CGROUP_MAX_VAL];
+
+ mounts = fopen("/proc/mounts", "r");
+ if (mounts == NULL) {
+ VIR_ERROR(_("Unable to open /proc/mounts"));
+ return -ENOENT;
+ }
+
+ while (getmntent_r(mounts, &entry, buf, sizeof(buf)) != NULL) {
+ if (STRNEQ(entry.mnt_type, "cgroup"))
+ continue;
+
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ const char *typestr = virCgroupControllerTypeToString(i);
+ int typelen = strlen(typestr);
+ char *tmp = entry.mnt_opts;
+ while (tmp) {
+ char *next = strchr(tmp, ',');
+ int len;
+ if (next) {
+ len = next-tmp;
+ next++;
+ } else {
+ len = strlen(tmp);
+ }
+ /* NB, the same controller can appear >1 time in mount list
+ * due to bind mounts from one location to another. Pick the
+ * first entry only
+ */
+ if (typelen == len && STREQLEN(typestr, tmp, len) &&
+ !group->controllers[i].mountPoint &&
+ !(group->controllers[i].mountPoint = strdup(entry.mnt_dir)))
+ goto no_memory;
+ tmp = next;
+ }
+ }
+ }
+
+ VIR_FORCE_FCLOSE(mounts);
+
+ return 0;
+
+no_memory:
+ VIR_FORCE_FCLOSE(mounts);
+ return -ENOMEM;
+}
+
+
+/*
+ * Process /proc/self/cgroup figuring out what cgroup
+ * sub-path the current process is assigned to. ie not
+ * necessarily in the root
+ */
+static int virCgroupDetectPlacement(virCgroupPtr group)
+{
+ int i;
+ FILE *mapping = NULL;
+ char line[1024];
+
+ mapping = fopen("/proc/self/cgroup", "r");
+ if (mapping == NULL) {
+ VIR_ERROR(_("Unable to open /proc/self/cgroup"));
+ return -ENOENT;
+ }
+
+ while (fgets(line, sizeof(line), mapping) != NULL) {
+ char *controllers = strchr(line, ':');
+ char *path = controllers ? strchr(controllers+1, ':') : NULL;
+ char *nl = path ? strchr(path, '\n') : NULL;
+
+ if (!controllers || !path)
+ continue;
+
+ if (nl)
+ *nl = '\0';
+
+ *path = '\0';
+ controllers++;
+ path++;
+
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ const char *typestr = virCgroupControllerTypeToString(i);
+ int typelen = strlen(typestr);
+ char *tmp = controllers;
+ while (tmp) {
+ char *next = strchr(tmp, ',');
+ int len;
+ if (next) {
+ len = next-tmp;
+ next++;
+ } else {
+ len = strlen(tmp);
+ }
+ if (typelen == len && STREQLEN(typestr, tmp, len) &&
+ !(group->controllers[i].placement = strdup(STREQ(path, "/") ? "" : path)))
+ goto no_memory;
+
+ tmp = next;
+ }
+ }
+ }
+
+ VIR_FORCE_FCLOSE(mapping);
+
+ return 0;
+
+no_memory:
+ VIR_FORCE_FCLOSE(mapping);
+ return -ENOMEM;
+
+}
+
+static int virCgroupDetect(virCgroupPtr group)
+{
+ int any = 0;
+ int rc;
+ int i;
+
+ rc = virCgroupDetectMounts(group);
+ if (rc < 0) {
+ VIR_ERROR(_("Failed to detect mounts for %s"), group->path);
+ return rc;
+ }
+
+ /* Check that at least 1 controller is available */
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ if (group->controllers[i].mountPoint != NULL)
+ any = 1;
+ }
+ if (!any)
+ return -ENXIO;
+
+
+ rc = virCgroupDetectPlacement(group);
+
+ if (rc == 0) {
+ /* Check that for every mounted controller, we found our placement */
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ if (!group->controllers[i].mountPoint)
+ continue;
+
+ if (!group->controllers[i].placement) {
+ VIR_ERROR(_("Could not find placement for controller %s at %s"),
+ virCgroupControllerTypeToString(i),
+ group->controllers[i].placement);
+ rc = -ENOENT;
+ break;
+ }
+
+ VIR_DEBUG("Detected mount/mapping %i:%s at %s in %s", i,
+ virCgroupControllerTypeToString(i),
+ group->controllers[i].mountPoint,
+ group->controllers[i].placement);
+ }
+ } else {
+ VIR_ERROR(_("Failed to detect mapping for %s"), group->path);
+ }
+
+ return rc;
+}
+#endif
+
+
+int virCgroupPathOfController(virCgroupPtr group,
+ int controller,
+ const char *key,
+ char **path)
+{
+ if (controller == -1) {
+ int i;
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ if (group->controllers[i].mountPoint &&
+ group->controllers[i].placement) {
+ controller = i;
+ break;
+ }
+ }
+ }
+ if (controller == -1)
+ return -ENOSYS;
+
+ if (group->controllers[controller].mountPoint == NULL)
+ return -ENOENT;
+
+ if (group->controllers[controller].placement == NULL)
+ return -ENOENT;
+
+ if (virAsprintf(path, "%s%s%s/%s",
+ group->controllers[controller].mountPoint,
+ group->controllers[controller].placement,
+ STREQ(group->path, "/") ? "" : group->path,
+ key ? key : "") == -1)
+ return -ENOMEM;
+
+ return 0;
+}
+
+
+static int virCgroupSetValueStr(virCgroupPtr group,
+ int controller,
+ const char *key,
+ const char *value)
+{
+ int rc = 0;
+ char *keypath = NULL;
+
+ rc = virCgroupPathOfController(group, controller, key, &keypath);
+ if (rc != 0)
+ return rc;
+
+ VIR_DEBUG("Set value '%s' to '%s'", keypath, value);
+ rc = virFileWriteStr(keypath, value, 0);
+ if (rc < 0) {
+ rc = -errno;
+ VIR_DEBUG("Failed to write value '%s': %m", value);
+ } else {
+ rc = 0;
+ }
+
+ VIR_FREE(keypath);
+
+ return rc;
+}
+
+static int virCgroupGetValueStr(virCgroupPtr group,
+ int controller,
+ const char *key,
+ char **value)
+{
+ int rc;
+ char *keypath = NULL;
+
+ *value = NULL;
+
+ rc = virCgroupPathOfController(group, controller, key, &keypath);
+ if (rc != 0) {
+ VIR_DEBUG("No path of %s, %s", group->path, key);
+ return rc;
+ }
+
+ VIR_DEBUG("Get value %s", keypath);
+
+ rc = virFileReadAll(keypath, 1024*1024, value);
+ if (rc < 0) {
+ rc = -errno;
+ VIR_DEBUG("Failed to read %s: %m\n", keypath);
+ } else {
+ /* Terminated with '\n' has sometimes harmful effects to the caller */
+ if ((*value)[rc - 1] == '\n')
+ (*value)[rc - 1] = '\0';
+
+ rc = 0;
+ }
+
+ VIR_FREE(keypath);
+
+ return rc;
+}
+
+static int virCgroupSetValueU64(virCgroupPtr group,
+ int controller,
+ const char *key,
+ unsigned long long int value)
+{
+ char *strval = NULL;
+ int rc;
+
+ if (virAsprintf(&strval, "%llu", value) == -1)
+ return -ENOMEM;
+
+ rc = virCgroupSetValueStr(group, controller, key, strval);
+
+ VIR_FREE(strval);
+
+ return rc;
+}
+
+
+
+static int virCgroupSetValueI64(virCgroupPtr group,
+ int controller,
+ const char *key,
+ long long int value)
+{
+ char *strval = NULL;
+ int rc;
+
+ if (virAsprintf(&strval, "%lld", value) == -1)
+ return -ENOMEM;
+
+ rc = virCgroupSetValueStr(group, controller, key, strval);
+
+ VIR_FREE(strval);
+
+ return rc;
+}
+
+static int virCgroupGetValueI64(virCgroupPtr group,
+ int controller,
+ const char *key,
+ long long int *value)
+{
+ char *strval = NULL;
+ int rc = 0;
+
+ rc = virCgroupGetValueStr(group, controller, key, &strval);
+ if (rc != 0)
+ goto out;
+
+ if (virStrToLong_ll(strval, NULL, 10, value) < 0)
+ rc = -EINVAL;
+out:
+ VIR_FREE(strval);
+
+ return rc;
+}
+
+static int virCgroupGetValueU64(virCgroupPtr group,
+ int controller,
+ const char *key,
+ unsigned long long int *value)
+{
+ char *strval = NULL;
+ int rc = 0;
+
+ rc = virCgroupGetValueStr(group, controller, key, &strval);
+ if (rc != 0)
+ goto out;
+
+ if (virStrToLong_ull(strval, NULL, 10, value) < 0)
+ rc = -EINVAL;
+out:
+ VIR_FREE(strval);
+
+ return rc;
+}
+
+
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+static int virCgroupCpuSetInherit(virCgroupPtr parent, virCgroupPtr group)
+{
+ int i;
+ int rc = 0;
+ const char *inherit_values[] = {
+ "cpuset.cpus",
+ "cpuset.mems",
+ };
+
+ VIR_DEBUG("Setting up inheritance %s -> %s", parent->path, group->path);
+ for (i = 0; i < ARRAY_CARDINALITY(inherit_values) ; i++) {
+ char *value;
+
+ rc = virCgroupGetValueStr(parent,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ inherit_values[i],
+ &value);
+ if (rc != 0) {
+ VIR_ERROR(_("Failed to get %s %d"), inherit_values[i], rc);
+ break;
+ }
+
+ VIR_DEBUG("Inherit %s = %s", inherit_values[i], value);
+
+ rc = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ inherit_values[i],
+ value);
+ VIR_FREE(value);
+
+ if (rc != 0) {
+ VIR_ERROR(_("Failed to set %s %d"), inherit_values[i], rc);
+ break;
+ }
+ }
+
+ return rc;
+}
+
+static int virCgroupSetMemoryUseHierarchy(virCgroupPtr group)
+{
+ int rc = 0;
+ unsigned long long value;
+ const char *filename = "memory.use_hierarchy";
+
+ rc = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ filename, &value);
+ if (rc != 0) {
+ VIR_ERROR(_("Failed to read %s/%s (%d)"), group->path, filename, rc);
+ return rc;
+ }
+
+ /* Setting twice causes error, so if already enabled, skip setting */
+ if (value == 1)
+ return 0;
+
+ VIR_DEBUG("Setting up %s/%s", group->path, filename);
+ rc = virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ filename, 1);
+
+ if (rc != 0) {
+ VIR_ERROR(_("Failed to set %s/%s (%d)"), group->path, filename, rc);
+ }
+
+ return rc;
+}
+
+static int virCgroupMakeGroup(virCgroupPtr parent,
+ virCgroupPtr group,
+ bool create,
+ unsigned int flags)
+{
+ int i;
+ int rc = 0;
+
+ VIR_DEBUG("Make group %s", group->path);
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ char *path = NULL;
+
+ /* Skip over controllers that aren't mounted */
+ if (!group->controllers[i].mountPoint)
+ continue;
+
+ /* We need to control cpu bandwidth for each vcpu now */
+ if ((flags & VIR_CGROUP_VCPU) &&
+ (i != VIR_CGROUP_CONTROLLER_CPU &&
+ i != VIR_CGROUP_CONTROLLER_CPUACCT &&
+ i != VIR_CGROUP_CONTROLLER_CPUSET)) {
+ /* treat it as unmounted and we can use virCgroupAddTask */
+ VIR_FREE(group->controllers[i].mountPoint);
+ continue;
+ }
+
+ rc = virCgroupPathOfController(group, i, "", &path);
+ if (rc < 0)
+ return rc;
+ /* As of Feb 2011, clang can't see that the above function
+ * call did not modify group. */
+ sa_assert(group->controllers[i].mountPoint);
+
+ VIR_DEBUG("Make controller %s", path);
+ if (access(path, F_OK) != 0) {
+ if (!create ||
+ mkdir(path, 0755) < 0) {
+ /* With a kernel that doesn't support multi-level directory
+ * for blkio controller, libvirt will fail and disable all
+ * other controllers even though they are available. So
+ * treat blkio as unmounted if mkdir fails. */
+ if (i == VIR_CGROUP_CONTROLLER_BLKIO) {
+ rc = 0;
+ VIR_FREE(group->controllers[i].mountPoint);
+ VIR_FREE(path);
+ continue;
+ } else {
+ rc = -errno;
+ VIR_FREE(path);
+ break;
+ }
+ }
+ if (group->controllers[VIR_CGROUP_CONTROLLER_CPUSET].mountPoint != NULL &&
+ (i == VIR_CGROUP_CONTROLLER_CPUSET ||
+ STREQ(group->controllers[i].mountPoint, group->controllers[VIR_CGROUP_CONTROLLER_CPUSET].mountPoint))) {
+ rc = virCgroupCpuSetInherit(parent, group);
+ if (rc != 0) {
+ VIR_FREE(path);
+ break;
+ }
+ }
+ /*
+ * Note that virCgroupSetMemoryUseHierarchy should always be
+ * called prior to creating subcgroups and attaching tasks.
+ */
+ if ((flags & VIR_CGROUP_MEM_HIERACHY) &&
+ (group->controllers[VIR_CGROUP_CONTROLLER_MEMORY].mountPoint != NULL) &&
+ (i == VIR_CGROUP_CONTROLLER_MEMORY ||
+ STREQ(group->controllers[i].mountPoint, group->controllers[VIR_CGROUP_CONTROLLER_MEMORY].mountPoint))) {
+ rc = virCgroupSetMemoryUseHierarchy(group);
+ if (rc != 0) {
+ VIR_FREE(path);
+ break;
+ }
+ }
+ }
+
+ VIR_FREE(path);
+ }
+
+ return rc;
+}
+
+
+static int virCgroupNew(const char *path,
+ virCgroupPtr *group)
+{
+ int rc = 0;
+ char *typpath = NULL;
+
+ VIR_DEBUG("New group %s", path);
+ *group = NULL;
+
+ if (VIR_ALLOC((*group)) != 0) {
+ rc = -ENOMEM;
+ goto err;
+ }
+
+ if (!((*group)->path = strdup(path))) {
+ rc = -ENOMEM;
+ goto err;
+ }
+
+ rc = virCgroupDetect(*group);
+ if (rc < 0)
+ goto err;
+
+ return rc;
+err:
+ virCgroupFree(group);
+ *group = NULL;
+
+ VIR_FREE(typpath);
+
+ return rc;
+}
+
+static int virCgroupAppRoot(bool privileged,
+ virCgroupPtr *group,
+ bool create)
+{
+ virCgroupPtr rootgrp = NULL;
+ int rc;
+
+ rc = virCgroupNew("/", &rootgrp);
+ if (rc != 0)
+ return rc;
+
+ if (privileged) {
+ rc = virCgroupNew("/libvirt", group);
+ } else {
+ char *rootname;
+ char *username;
+ username = virGetUserName(getuid());
+ if (!username) {
+ rc = -ENOMEM;
+ goto cleanup;
+ }
+ rc = virAsprintf(&rootname, "/libvirt-%s", username);
+ VIR_FREE(username);
+ if (rc < 0) {
+ rc = -ENOMEM;
+ goto cleanup;
+ }
+
+ rc = virCgroupNew(rootname, group);
+ VIR_FREE(rootname);
+ }
+ if (rc != 0)
+ goto cleanup;
+
+ rc = virCgroupMakeGroup(rootgrp, *group, create, VIR_CGROUP_NONE);
+
+cleanup:
+ virCgroupFree(&rootgrp);
+ return rc;
+}
+#endif
+
+#if defined _DIRENT_HAVE_D_TYPE
+static int virCgroupRemoveRecursively(char *grppath)
+{
+ DIR *grpdir;
+ struct dirent *ent;
+ int rc = 0;
+
+ grpdir = opendir(grppath);
+ if (grpdir == NULL) {
+ if (errno == ENOENT)
+ return 0;
+ rc = -errno;
+ VIR_ERROR(_("Unable to open %s (%d)"), grppath, errno);
+ return rc;
+ }
+
+ for (;;) {
+ char *path;
+
+ errno = 0;
+ ent = readdir(grpdir);
+ if (ent == NULL) {
+ if ((rc = -errno))
+ VIR_ERROR(_("Failed to readdir for %s (%d)"), grppath, errno);
+ break;
+ }
+
+ if (ent->d_name[0] == '.') continue;
+ if (ent->d_type != DT_DIR) continue;
+
+ if (virAsprintf(&path, "%s/%s", grppath, ent->d_name) == -1) {
+ rc = -ENOMEM;
+ break;
+ }
+ rc = virCgroupRemoveRecursively(path);
+ VIR_FREE(path);
+ if (rc != 0)
+ break;
+ }
+ closedir(grpdir);
+
+ VIR_DEBUG("Removing cgroup %s", grppath);
+ if (rmdir(grppath) != 0 && errno != ENOENT) {
+ rc = -errno;
+ VIR_ERROR(_("Unable to remove %s (%d)"), grppath, errno);
+ }
+
+ return rc;
+}
+#else
+static int virCgroupRemoveRecursively(char *grppath ATTRIBUTE_UNUSED)
+{
+ /* Claim no support */
+ return -ENXIO;
+}
+#endif
+
+/**
+ * virCgroupRemove:
+ *
+ * @group: The group to be removed
+ *
+ * It first removes all child groups recursively
+ * in depth first order and then removes @group
+ * because the presence of the child groups
+ * prevents removing @group.
+ *
+ * Returns: 0 on success
+ */
+int virCgroupRemove(virCgroupPtr group)
+{
+ int rc = 0;
+ int i;
+ char *grppath = NULL;
+
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ /* Skip over controllers not mounted */
+ if (!group->controllers[i].mountPoint)
+ continue;
+
+ if (virCgroupPathOfController(group,
+ i,
+ NULL,
+ &grppath) != 0)
+ continue;
+
+ VIR_DEBUG("Removing cgroup %s and all child cgroups", grppath);
+ rc = virCgroupRemoveRecursively(grppath);
+ VIR_FREE(grppath);
+ }
+
+ return rc;
+}
+
+/**
+ * virCgroupAddTask:
+ *
+ * @group: The cgroup to add a task to
+ * @pid: The pid of the task to add
+ *
+ * Returns: 0 on success
+ */
+int virCgroupAddTask(virCgroupPtr group, pid_t pid)
+{
+ int rc = 0;
+ int i;
+
+ for (i = 0 ; i < VIR_CGROUP_CONTROLLER_LAST ; i++) {
+ /* Skip over controllers not mounted */
+ if (!group->controllers[i].mountPoint)
+ continue;
+
+ rc = virCgroupSetValueU64(group, i, "tasks", (unsigned long long)pid);
+ if (rc != 0)
+ break;
+ }
+
+ return rc;
+}
+
+/**
+ * virCgroupAddTaskController:
+ *
+ * @group: The cgroup to add a task to
+ * @pid: The pid of the task to add
+ * @controller: The cgroup controller to be operated on
+ *
+ * Returns: 0 on success or -errno on failure
+ */
+int virCgroupAddTaskController(virCgroupPtr group, pid_t pid, int controller)
+{
+ if (controller < 0 || controller >= VIR_CGROUP_CONTROLLER_LAST)
+ return -EINVAL;
+
+ if (!group->controllers[controller].mountPoint)
+ return -EINVAL;
+
+ return virCgroupSetValueU64(group, controller, "tasks",
+ (unsigned long long)pid);
+}
+
+
+static int virCgroupAddTaskStrController(virCgroupPtr group,
+ const char *pidstr,
+ int controller)
+{
+ char *str = NULL, *cur = NULL, *next = NULL;
+ unsigned long long p = 0;
+ int rc = 0;
+ char *endp;
+
+ if (virAsprintf(&str, "%s", pidstr) < 0)
+ return -1;
+
+ cur = str;
+ while (*cur != '\0') {
+ rc = virStrToLong_ull(cur, &endp, 10, &p);
+ if (rc != 0)
+ goto cleanup;
+
+ rc = virCgroupAddTaskController(group, p, controller);
+ if (rc != 0)
+ goto cleanup;
+
+ next = strchr(cur, '\n');
+ if (next) {
+ cur = next + 1;
+ *next = '\0';
+ } else {
+ break;
+ }
+ }
+
+cleanup:
+ VIR_FREE(str);
+ return rc;
+}
+
+/**
+ * virCgroupMoveTask:
+ *
+ * @src_group: The source cgroup where all tasks are removed from
+ * @dest_group: The destination where all tasks are added to
+ * @controller: The cgroup controller to be operated on
+ *
+ * Returns: 0 on success or -errno on failure
+ */
+int virCgroupMoveTask(virCgroupPtr src_group, virCgroupPtr dest_group,
+ int controller)
+{
+ int rc = 0, err = 0;
+ char *content = NULL;
+
+ if (controller < VIR_CGROUP_CONTROLLER_CPU ||
+ controller > VIR_CGROUP_CONTROLLER_BLKIO)
+ return -EINVAL;
+
+ if (!src_group->controllers[controller].mountPoint ||
+ !dest_group->controllers[controller].mountPoint) {
+ VIR_WARN("no vm cgroup in controller %d", controller);
+ return 0;
+ }
+
+ rc = virCgroupGetValueStr(src_group, controller, "tasks", &content);
+ if (rc != 0)
+ return rc;
+
+ rc = virCgroupAddTaskStrController(dest_group, content, controller);
+ if (rc != 0)
+ goto cleanup;
+
+ VIR_FREE(content);
+
+ return 0;
+
+cleanup:
+ /*
+ * We don't need to recover dest_cgroup because cgroup will make sure
+ * that one task only resides in one cgroup of the same controller.
+ */
+ err = virCgroupAddTaskStrController(src_group, content, controller);
+ if (err != 0)
+ VIR_ERROR(_("Cannot recover cgroup %s from %s"),
+ src_group->controllers[controller].mountPoint,
+ dest_group->controllers[controller].mountPoint);
+ VIR_FREE(content);
+
+ return rc;
+}
+
+/**
+ * virCgroupForDriver:
+ *
+ * @name: name of this driver (e.g., xen, qemu, lxc)
+ * @group: Pointer to returned virCgroupPtr
+ *
+ * Returns 0 on success
+ */
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+int virCgroupForDriver(const char *name,
+ virCgroupPtr *group,
+ bool privileged,
+ bool create)
+{
+ int rc;
+ char *path = NULL;
+ virCgroupPtr rootgrp = NULL;
+
+ rc = virCgroupAppRoot(privileged, &rootgrp, create);
+ if (rc != 0)
+ goto out;
+
+ if (virAsprintf(&path, "%s/%s", rootgrp->path, name) < 0) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = virCgroupNew(path, group);
+ VIR_FREE(path);
+
+ if (rc == 0) {
+ rc = virCgroupMakeGroup(rootgrp, *group, create, VIR_CGROUP_NONE);
+ if (rc != 0)
+ virCgroupFree(group);
+ }
+
+out:
+ virCgroupFree(&rootgrp);
+
+ return rc;
+}
+#else
+int virCgroupForDriver(const char *name ATTRIBUTE_UNUSED,
+ virCgroupPtr *group ATTRIBUTE_UNUSED,
+ bool privileged ATTRIBUTE_UNUSED,
+ bool create ATTRIBUTE_UNUSED)
+{
+ /* Claim no support */
+ return -ENXIO;
+}
+#endif
+
+/**
+* virCgroupGetAppRoot:
+*
+* @group: Pointer to returned virCgroupPtr
+*
+* Returns 0 on success
+*/
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+int virCgroupGetAppRoot(virCgroupPtr *group)
+{
+ return virCgroupNew("/", group);
+}
+#else
+int virCgroupGetAppRoot(virCgroupPtr *group ATTRIBUTE_UNUSED)
+{
+ return -ENXIO;
+}
+#endif
+
+/**
+ * virCgroupForDomain:
+ *
+ * @driver: group for driver owning the domain
+ * @name: name of the domain
+ * @group: Pointer to returned virCgroupPtr
+ *
+ * Returns 0 on success
+ */
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+int virCgroupForDomain(virCgroupPtr driver,
+ const char *name,
+ virCgroupPtr *group,
+ bool create)
+{
+ int rc;
+ char *path;
+
+ if (driver == NULL)
+ return -EINVAL;
+
+ if (virAsprintf(&path, "%s/%s", driver->path, name) < 0)
+ return -ENOMEM;
+
+ rc = virCgroupNew(path, group);
+ VIR_FREE(path);
+
+ if (rc == 0) {
+ /*
+ * Create a cgroup with memory.use_hierarchy enabled to
+ * surely account memory usage of lxc with ns subsystem
+ * enabled. (To be exact, memory and ns subsystems are
+ * enabled at the same time.)
+ *
+ * The reason why doing it here, not a upper group, say
+ * a group for driver, is to avoid overhead to track
+ * cumulative usage that we don't need.
+ */
+ rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_MEM_HIERACHY);
+ if (rc != 0)
+ virCgroupFree(group);
+ }
+
+ return rc;
+}
+#else
+int virCgroupForDomain(virCgroupPtr driver ATTRIBUTE_UNUSED,
+ const char *name ATTRIBUTE_UNUSED,
+ virCgroupPtr *group ATTRIBUTE_UNUSED,
+ bool create ATTRIBUTE_UNUSED)
+{
+ return -ENXIO;
+}
+#endif
+
+/**
+ * virCgroupForVcpu:
+ *
+ * @driver: group for the domain
+ * @vcpuid: id of the vcpu
+ * @group: Pointer to returned virCgroupPtr
+ *
+ * Returns 0 on success
+ */
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+int virCgroupForVcpu(virCgroupPtr driver,
+ int vcpuid,
+ virCgroupPtr *group,
+ bool create)
+{
+ int rc;
+ char *path;
+
+ if (driver == NULL)
+ return -EINVAL;
+
+ if (virAsprintf(&path, "%s/vcpu%d", driver->path, vcpuid) < 0)
+ return -ENOMEM;
+
+ rc = virCgroupNew(path, group);
+ VIR_FREE(path);
+
+ if (rc == 0) {
+ rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_VCPU);
+ if (rc != 0)
+ virCgroupFree(group);
+ }
+
+ return rc;
+}
+#else
+int virCgroupForVcpu(virCgroupPtr driver ATTRIBUTE_UNUSED,
+ int vcpuid ATTRIBUTE_UNUSED,
+ virCgroupPtr *group ATTRIBUTE_UNUSED,
+ bool create ATTRIBUTE_UNUSED)
+{
+ return -ENXIO;
+}
+#endif
+
+/**
+ * virCgroupForEmulator:
+ *
+ * @driver: group for the domain
+ * @group: Pointer to returned virCgroupPtr
+ *
+ * Returns: 0 on success or -errno on failure
+ */
+#if defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+int virCgroupForEmulator(virCgroupPtr driver,
+ virCgroupPtr *group,
+ bool create)
+{
+ int rc;
+ char *path;
+
+ if (driver == NULL)
+ return -EINVAL;
+
+ if (virAsprintf(&path, "%s/emulator", driver->path) < 0)
+ return -ENOMEM;
+
+ rc = virCgroupNew(path, group);
+ VIR_FREE(path);
+
+ if (rc == 0) {
+ rc = virCgroupMakeGroup(driver, *group, create, VIR_CGROUP_VCPU);
+ if (rc != 0)
+ virCgroupFree(group);
+ }
+
+ return rc;
+}
+#else
+int virCgroupForEmulator(virCgroupPtr driver ATTRIBUTE_UNUSED,
+ virCgroupPtr *group ATTRIBUTE_UNUSED,
+ bool create ATTRIBUTE_UNUSED)
+{
+ return -ENXIO;
+}
+
+#endif
+/**
+ * virCgroupSetBlkioWeight:
+ *
+ * @group: The cgroup to change io weight for
+ * @weight: The Weight for this cgroup
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetBlkioWeight(virCgroupPtr group, unsigned int weight)
+{
+ if (weight > 1000 || weight < 100)
+ return -EINVAL;
+
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_BLKIO,
+ "blkio.weight",
+ weight);
+}
+
+/**
+ * virCgroupGetBlkioWeight:
+ *
+ * @group: The cgroup to get weight for
+ * @Weight: Pointer to returned weight
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetBlkioWeight(virCgroupPtr group, unsigned int *weight)
+{
+ unsigned long long tmp;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_BLKIO,
+ "blkio.weight", &tmp);
+ if (ret == 0)
+ *weight = tmp;
+ return ret;
+}
+
+/**
+ * virCgroupSetBlkioDeviceWeight:
+ *
+ * @group: The cgroup to change io device weight device for
+ * @path: The device with a weight to alter
+ * @weight: The new device weight (100-1000), or 0 to clear
+ *
+ * device_weight is treated as a write-only parameter, so
+ * there isn't a getter counterpart.
+ *
+ * Returns: 0 on success, -errno on failure
+ */
+#if defined(major) && defined(minor)
+int virCgroupSetBlkioDeviceWeight(virCgroupPtr group,
+ const char *path,
+ unsigned int weight)
+{
+ char *str;
+ struct stat sb;
+ int ret;
+
+ if (weight && (weight > 1000 || weight < 100))
+ return -EINVAL;
+
+ if (stat(path, &sb) < 0)
+ return -errno;
+
+ if (!S_ISBLK(sb.st_mode))
+ return -EINVAL;
+
+ if (virAsprintf(&str, "%d:%d %d", major(sb.st_rdev), minor(sb.st_rdev),
+ weight) < 0)
+ return -errno;
+
+ ret = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_BLKIO,
+ "blkio.weight_device",
+ str);
+ VIR_FREE(str);
+ return ret;
+}
+#else
+int
+virCgroupSetBlkioDeviceWeight(virCgroupPtr group ATTRIBUTE_UNUSED,
+ const char *path ATTRIBUTE_UNUSED,
+ unsigned int weight ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+#endif
+
+/**
+ * virCgroupSetMemory:
+ *
+ * @group: The cgroup to change memory for
+ * @kb: The memory amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetMemory(virCgroupPtr group, unsigned long long kb)
+{
+ unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
+
+ if (kb > maxkb)
+ return -EINVAL;
+ else if (kb == maxkb)
+ return virCgroupSetValueI64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.limit_in_bytes",
+ -1);
+ else
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.limit_in_bytes",
+ kb << 10);
+}
+
+/**
+ * virCgroupGetMemoryUsage:
+ *
+ * @group: The cgroup to change memory for
+ * @kb: Pointer to returned used memory in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetMemoryUsage(virCgroupPtr group, unsigned long *kb)
+{
+ long long unsigned int usage_in_bytes;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.usage_in_bytes", &usage_in_bytes);
+ if (ret == 0)
+ *kb = (unsigned long) usage_in_bytes >> 10;
+ return ret;
+}
+
+/**
+ * virCgroupSetMemoryHardLimit:
+ *
+ * @group: The cgroup to change memory hard limit for
+ * @kb: The memory amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetMemoryHardLimit(virCgroupPtr group, unsigned long long kb)
+{
+ return virCgroupSetMemory(group, kb);
+}
+
+/**
+ * virCgroupGetMemoryHardLimit:
+ *
+ * @group: The cgroup to get the memory hard limit for
+ * @kb: The memory amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetMemoryHardLimit(virCgroupPtr group, unsigned long long *kb)
+{
+ long long unsigned int limit_in_bytes;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.limit_in_bytes", &limit_in_bytes);
+ if (ret == 0)
+ *kb = limit_in_bytes >> 10;
+ return ret;
+}
+
+/**
+ * virCgroupSetMemorySoftLimit:
+ *
+ * @group: The cgroup to change memory soft limit for
+ * @kb: The memory amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetMemorySoftLimit(virCgroupPtr group, unsigned long long kb)
+{
+ unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
+
+ if (kb > maxkb)
+ return -EINVAL;
+ else if (kb == maxkb)
+ return virCgroupSetValueI64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.soft_limit_in_bytes",
+ -1);
+ else
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.soft_limit_in_bytes",
+ kb << 10);
+}
+
+
+/**
+ * virCgroupGetMemorySoftLimit:
+ *
+ * @group: The cgroup to get the memory soft limit for
+ * @kb: The memory amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetMemorySoftLimit(virCgroupPtr group, unsigned long long *kb)
+{
+ long long unsigned int limit_in_bytes;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.soft_limit_in_bytes", &limit_in_bytes);
+ if (ret == 0)
+ *kb = limit_in_bytes >> 10;
+ return ret;
+}
+
+/**
+ * virCgroupSetMemSwapHardLimit:
+ *
+ * @group: The cgroup to change mem+swap hard limit for
+ * @kb: The mem+swap amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetMemSwapHardLimit(virCgroupPtr group, unsigned long long kb)
+{
+ unsigned long long maxkb = VIR_DOMAIN_MEMORY_PARAM_UNLIMITED;
+
+ if (kb > maxkb)
+ return -EINVAL;
+ else if (kb == maxkb)
+ return virCgroupSetValueI64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.memsw.limit_in_bytes",
+ -1);
+ else
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.memsw.limit_in_bytes",
+ kb << 10);
+}
+
+/**
+ * virCgroupGetMemSwapHardLimit:
+ *
+ * @group: The cgroup to get mem+swap hard limit for
+ * @kb: The mem+swap amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetMemSwapHardLimit(virCgroupPtr group, unsigned long long *kb)
+{
+ long long unsigned int limit_in_bytes;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.memsw.limit_in_bytes", &limit_in_bytes);
+ if (ret == 0)
+ *kb = limit_in_bytes >> 10;
+ return ret;
+}
+
+/**
+ * virCgroupGetMemSwapUsage:
+ *
+ * @group: The cgroup to get mem+swap usage for
+ * @kb: The mem+swap amount in kilobytes
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetMemSwapUsage(virCgroupPtr group, unsigned long long *kb)
+{
+ long long unsigned int usage_in_bytes;
+ int ret;
+ ret = virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ "memory.memsw.usage_in_bytes", &usage_in_bytes);
+ if (ret == 0)
+ *kb = usage_in_bytes >> 10;
+ return ret;
+}
+
+/**
+ * virCgroupSetCpusetMems:
+ *
+ * @group: The cgroup to set cpuset.mems for
+ * @mems: the numa nodes to set
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetCpusetMems(virCgroupPtr group, const char *mems)
+{
+ return virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ "cpuset.mems",
+ mems);
+}
+
+/**
+ * virCgroupGetCpusetMems:
+ *
+ * @group: The cgroup to get cpuset.mems for
+ * @mems: the numa nodes to get
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetCpusetMems(virCgroupPtr group, char **mems)
+{
+ return virCgroupGetValueStr(group,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ "cpuset.mems",
+ mems);
+}
+
+/**
+ * virCgroupSetCpusetCpus:
+ *
+ * @group: The cgroup to set cpuset.cpus for
+ * @cpus: the cpus to set
+ *
+ * Retuens: 0 on success
+ */
+int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus)
+{
+ return virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ "cpuset.cpus",
+ cpus);
+}
+
+/**
+ * virCgroupGetCpusetCpus:
+ *
+ * @group: The cgroup to get cpuset.cpus for
+ * @cpus: the cpus to get
+ *
+ * Retuens: 0 on success
+ */
+int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus)
+{
+ return virCgroupGetValueStr(group,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ "cpuset.cpus",
+ cpus);
+}
+
+/**
+ * virCgroupDenyAllDevices:
+ *
+ * @group: The cgroup to deny all permissions, for all devices
+ *
+ * Returns: 0 on success
+ */
+int virCgroupDenyAllDevices(virCgroupPtr group)
+{
+ return virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ "devices.deny",
+ "a");
+}
+
+/**
+ * virCgroupAllowDevice:
+ *
+ * @group: The cgroup to allow a device for
+ * @type: The device type (i.e., 'c' or 'b')
+ * @major: The major number of the device
+ * @minor: The minor number of the device
+ * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
+ *
+ * Returns: 0 on success
+ */
+int virCgroupAllowDevice(virCgroupPtr group, char type, int major, int minor,
+ int perms)
+{
+ int rc;
+ char *devstr = NULL;
+
+ if (virAsprintf(&devstr, "%c %i:%i %s%s%s", type, major, minor,
+ perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
+ perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
+ perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ "devices.allow",
+ devstr);
+out:
+ VIR_FREE(devstr);
+
+ return rc;
+}
+
+/**
+ * virCgroupAllowDeviceMajor:
+ *
+ * @group: The cgroup to allow an entire device major type for
+ * @type: The device type (i.e., 'c' or 'b')
+ * @major: The major number of the device type
+ * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
+ *
+ * Returns: 0 on success
+ */
+int virCgroupAllowDeviceMajor(virCgroupPtr group, char type, int major,
+ int perms)
+{
+ int rc;
+ char *devstr = NULL;
+
+ if (virAsprintf(&devstr, "%c %i:* %s%s%s", type, major,
+ perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
+ perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
+ perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ "devices.allow",
+ devstr);
+ out:
+ VIR_FREE(devstr);
+
+ return rc;
+}
+
+/**
+ * virCgroupAllowDevicePath:
+ *
+ * @group: The cgroup to allow the device for
+ * @path: the device to allow
+ * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to allow
+ *
+ * Queries the type of device and its major/minor number, and
+ * adds that to the cgroup ACL
+ *
+ * Returns: 0 on success, 1 if path exists but is not a device, or
+ * negative errno value on failure
+ */
+#if defined(major) && defined(minor)
+int virCgroupAllowDevicePath(virCgroupPtr group, const char *path, int perms)
+{
+ struct stat sb;
+
+ if (stat(path, &sb) < 0)
+ return -errno;
+
+ if (!S_ISCHR(sb.st_mode) && !S_ISBLK(sb.st_mode))
+ return 1;
+
+ return virCgroupAllowDevice(group,
+ S_ISCHR(sb.st_mode) ? 'c' : 'b',
+ major(sb.st_rdev),
+ minor(sb.st_rdev),
+ perms);
+}
+#else
+int virCgroupAllowDevicePath(virCgroupPtr group ATTRIBUTE_UNUSED,
+ const char *path ATTRIBUTE_UNUSED,
+ int perms ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+#endif
+
+
+/**
+ * virCgroupDenyDevice:
+ *
+ * @group: The cgroup to deny a device for
+ * @type: The device type (i.e., 'c' or 'b')
+ * @major: The major number of the device
+ * @minor: The minor number of the device
+ * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to deny
+ *
+ * Returns: 0 on success
+ */
+int virCgroupDenyDevice(virCgroupPtr group, char type, int major, int minor,
+ int perms)
+{
+ int rc;
+ char *devstr = NULL;
+
+ if (virAsprintf(&devstr, "%c %i:%i %s%s%s", type, major, minor,
+ perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
+ perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
+ perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ "devices.deny",
+ devstr);
+out:
+ VIR_FREE(devstr);
+
+ return rc;
+}
+
+/**
+ * virCgroupDenyDeviceMajor:
+ *
+ * @group: The cgroup to deny an entire device major type for
+ * @type: The device type (i.e., 'c' or 'b')
+ * @major: The major number of the device type
+ * @perms: Bitwise or of VIR_CGROUP_DEVICE permission bits to deny
+ *
+ * Returns: 0 on success
+ */
+int virCgroupDenyDeviceMajor(virCgroupPtr group, char type, int major,
+ int perms)
+{
+ int rc;
+ char *devstr = NULL;
+
+ if (virAsprintf(&devstr, "%c %i:* %s%s%s", type, major,
+ perms & VIR_CGROUP_DEVICE_READ ? "r" : "",
+ perms & VIR_CGROUP_DEVICE_WRITE ? "w" : "",
+ perms & VIR_CGROUP_DEVICE_MKNOD ? "m" : "") == -1) {
+ rc = -ENOMEM;
+ goto out;
+ }
+
+ rc = virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ "devices.deny",
+ devstr);
+ out:
+ VIR_FREE(devstr);
+
+ return rc;
+}
+
+#if defined(major) && defined(minor)
+int virCgroupDenyDevicePath(virCgroupPtr group, const char *path, int perms)
+{
+ struct stat sb;
+
+ if (stat(path, &sb) < 0)
+ return -errno;
+
+ if (!S_ISCHR(sb.st_mode) && !S_ISBLK(sb.st_mode))
+ return 1;
+
+ return virCgroupDenyDevice(group,
+ S_ISCHR(sb.st_mode) ? 'c' : 'b',
+ major(sb.st_rdev),
+ minor(sb.st_rdev),
+ perms);
+}
+#else
+int virCgroupDenyDevicePath(virCgroupPtr group ATTRIBUTE_UNUSED,
+ const char *path ATTRIBUTE_UNUSED,
+ int perms ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+#endif
+
+int virCgroupSetCpuShares(virCgroupPtr group, unsigned long long shares)
+{
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.shares", shares);
+}
+
+int virCgroupGetCpuShares(virCgroupPtr group, unsigned long long *shares)
+{
+ return virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.shares", shares);
+}
+
+/**
+ * virCgroupSetCpuCfsPeriod:
+ *
+ * @group: The cgroup to change cpu.cfs_period_us for
+ * @cfs_period: The bandwidth period in usecs
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetCpuCfsPeriod(virCgroupPtr group, unsigned long long cfs_period)
+{
+ /* The cfs_period shoule be greater or equal than 1ms, and less or equal
+ * than 1s.
+ */
+ if (cfs_period < 1000 || cfs_period > 1000000)
+ return -EINVAL;
+
+ return virCgroupSetValueU64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.cfs_period_us", cfs_period);
+}
+
+/**
+ * virCgroupGetCpuCfsPeriod:
+ *
+ * @group: The cgroup to get cpu.cfs_period_us for
+ * @cfs_period: Pointer to the returned bandwidth period in usecs
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetCpuCfsPeriod(virCgroupPtr group, unsigned long long *cfs_period)
+{
+ return virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.cfs_period_us", cfs_period);
+}
+
+/**
+ * virCgroupSetCpuCfsQuota:
+ *
+ * @group: The cgroup to change cpu.cfs_quota_us for
+ * @cfs_quota: the cpu bandwidth (in usecs) that this tg will be allowed to
+ * consume over period
+ *
+ * Returns: 0 on success
+ */
+int virCgroupSetCpuCfsQuota(virCgroupPtr group, long long cfs_quota)
+{
+ if (cfs_quota >= 0) {
+ /* The cfs_quota shoule be greater or equal than 1ms */
+ if (cfs_quota < 1000)
+ return -EINVAL;
+
+ /* check overflow */
+ if (cfs_quota > ULLONG_MAX / 1000)
+ return -EINVAL;
+ }
+
+ return virCgroupSetValueI64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.cfs_quota_us", cfs_quota);
+}
+
+/**
+ * virCgroupGetCpuCfsQuota:
+ *
+ * @group: The cgroup to get cpu.cfs_quota_us for
+ * @cfs_quota: Pointer to the returned cpu bandwidth (in usecs) that this tg
+ * will be allowed to consume over period
+ *
+ * Returns: 0 on success
+ */
+int virCgroupGetCpuCfsQuota(virCgroupPtr group, long long *cfs_quota)
+{
+ return virCgroupGetValueI64(group,
+ VIR_CGROUP_CONTROLLER_CPU,
+ "cpu.cfs_quota_us", cfs_quota);
+}
+
+int virCgroupGetCpuacctUsage(virCgroupPtr group, unsigned long long *usage)
+{
+ return virCgroupGetValueU64(group,
+ VIR_CGROUP_CONTROLLER_CPUACCT,
+ "cpuacct.usage", usage);
+}
+
+int virCgroupGetCpuacctPercpuUsage(virCgroupPtr group, char **usage)
+{
+ return virCgroupGetValueStr(group, VIR_CGROUP_CONTROLLER_CPUACCT,
+ "cpuacct.usage_percpu", usage);
+}
+
+#ifdef _SC_CLK_TCK
+int virCgroupGetCpuacctStat(virCgroupPtr group, unsigned long long *user,
+ unsigned long long *sys)
+{
+ char *str;
+ char *p;
+ int ret;
+ static double scale = -1.0;
+
+ if ((ret = virCgroupGetValueStr(group, VIR_CGROUP_CONTROLLER_CPUACCT,
+ "cpuacct.stat", &str)) < 0)
+ return ret;
+ if (!(p = STRSKIP(str, "user ")) ||
+ virStrToLong_ull(p, &p, 10, user) < 0 ||
+ !(p = STRSKIP(p, "\nsystem ")) ||
+ virStrToLong_ull(p, NULL, 10, sys) < 0) {
+ ret = -EINVAL;
+ goto cleanup;
+ }
+ /* times reported are in system ticks (generally 100 Hz), but that
+ * rate can theoretically vary between machines. Scale things
+ * into approximate nanoseconds. */
+ if (scale < 0) {
+ long ticks_per_sec = sysconf(_SC_CLK_TCK);
+ if (ticks_per_sec == -1) {
+ ret = -errno;
+ goto cleanup;
+ }
+ scale = 1000000000.0 / ticks_per_sec;
+ }
+ *user *= scale;
+ *sys *= scale;
+
+ ret = 0;
+cleanup:
+ VIR_FREE(str);
+ return ret;
+}
+#else
+int virCgroupGetCpuacctStat(virCgroupPtr group ATTRIBUTE_UNUSED,
+ unsigned long long *user ATTRIBUTE_UNUSED,
+ unsigned long long *sys ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+#endif
+
+int virCgroupSetFreezerState(virCgroupPtr group, const char *state)
+{
+ return virCgroupSetValueStr(group,
+ VIR_CGROUP_CONTROLLER_FREEZER,
+ "freezer.state", state);
+}
+
+int virCgroupGetFreezerState(virCgroupPtr group, char **state)
+{
+ return virCgroupGetValueStr(group,
+ VIR_CGROUP_CONTROLLER_FREEZER,
+ "freezer.state", state);
+}
+
+
+#if defined HAVE_KILL && defined HAVE_MNTENT_H && defined HAVE_GETMNTENT_R
+static int virCgroupKillInternal(virCgroupPtr group, int signum, virHashTablePtr pids)
+{
+ int rc;
+ int killedAny = 0;
+ char *keypath = NULL;
+ bool done = false;
+ FILE *fp = NULL;
+ VIR_DEBUG("group=%p path=%s signum=%d pids=%p",
+ group, group->path, signum, pids);
+
+ rc = virCgroupPathOfController(group, -1, "tasks", &keypath);
+ if (rc != 0) {
+ VIR_DEBUG("No path of %s, tasks", group->path);
+ return rc;
+ }
+
+ /* PIDs may be forking as we kill them, so loop
+ * until there are no new PIDs found
+ */
+ while (!done) {
+ done = true;
+ if (!(fp = fopen(keypath, "r"))) {
+ rc = -errno;
+ VIR_DEBUG("Failed to read %s: %m\n", keypath);
+ goto cleanup;
+ } else {
+ while (!feof(fp)) {
+ unsigned long pid_value;
+ if (fscanf(fp, "%lu", &pid_value) != 1) {
+ if (feof(fp))
+ break;
+ rc = -errno;
+ VIR_DEBUG("Failed to read %s: %m\n", keypath);
+ goto cleanup;
+ }
+ if (virHashLookup(pids, (void*)pid_value))
+ continue;
+
+ VIR_DEBUG("pid=%lu", pid_value);
+ /* Cgroups is a Linux concept, so this cast is safe. */
+ if (kill((pid_t)pid_value, signum) < 0) {
+ if (errno != ESRCH) {
+ rc = -errno;
+ goto cleanup;
+ }
+ /* Leave RC == 0 since we didn't kill one */
+ } else {
+ killedAny = 1;
+ done = false;
+ }
+
+ ignore_value(virHashAddEntry(pids, (void*)pid_value, (void*)1));
+ }
+ VIR_FORCE_FCLOSE(fp);
+ }
+ }
+
+ rc = killedAny ? 1 : 0;
+
+cleanup:
+ VIR_FREE(keypath);
+ VIR_FORCE_FCLOSE(fp);
+
+ return rc;
+}
+
+
+static uint32_t virCgroupPidCode(const void *name, uint32_t seed)
+{
+ unsigned long pid_value = (unsigned long)(intptr_t)name;
+ return virHashCodeGen(&pid_value, sizeof(pid_value), seed);
+}
+static bool virCgroupPidEqual(const void *namea, const void *nameb)
+{
+ return namea == nameb;
+}
+static void *virCgroupPidCopy(const void *name)
+{
+ return (void*)name;
+}
+
+/*
+ * Returns
+ * < 0 : errno that occurred
+ * 0 : no PIDs killed
+ * 1 : at least one PID killed
+ */
+int virCgroupKill(virCgroupPtr group, int signum)
+{
+ VIR_DEBUG("group=%p path=%s signum=%d", group, group->path, signum);
+ int rc;
+ /* The 'tasks' file in cgroups can contain duplicated
+ * pids, so we use a hash to track which we've already
+ * killed.
+ */
+ virHashTablePtr pids = virHashCreateFull(100,
+ NULL,
+ virCgroupPidCode,
+ virCgroupPidEqual,
+ virCgroupPidCopy,
+ NULL);
+
+ rc = virCgroupKillInternal(group, signum, pids);
+
+ virHashFree(pids);
+
+ return rc;
+}
+
+
+static int virCgroupKillRecursiveInternal(virCgroupPtr group, int signum, virHashTablePtr pids, bool dormdir)
+{
+ int rc;
+ int killedAny = 0;
+ char *keypath = NULL;
+ DIR *dp;
+ virCgroupPtr subgroup = NULL;
+ struct dirent *ent;
+ VIR_DEBUG("group=%p path=%s signum=%d pids=%p", group, group->path, signum, pids);
+
+ rc = virCgroupPathOfController(group, -1, "", &keypath);
+ if (rc != 0) {
+ VIR_DEBUG("No path of %s, tasks", group->path);
+ return rc;
+ }
+
+ if ((rc = virCgroupKillInternal(group, signum, pids)) != 0)
+ return rc;
+
+ VIR_DEBUG("Iterate over children of %s", keypath);
+ if (!(dp = opendir(keypath))) {
+ rc = -errno;
+ return rc;
+ }
+
+ while ((ent = readdir(dp))) {
+ char *subpath;
+
+ if (STREQ(ent->d_name, "."))
+ continue;
+ if (STREQ(ent->d_name, ".."))
+ continue;
+ if (ent->d_type != DT_DIR)
+ continue;
+
+ VIR_DEBUG("Process subdir %s", ent->d_name);
+ if (virAsprintf(&subpath, "%s/%s", group->path, ent->d_name) < 0) {
+ rc = -ENOMEM;
+ goto cleanup;
+ }
+
+ if ((rc = virCgroupNew(subpath, &subgroup)) != 0)
+ goto cleanup;
+
+ if ((rc = virCgroupKillRecursiveInternal(subgroup, signum, pids, true)) < 0)
+ goto cleanup;
+ if (rc == 1)
+ killedAny = 1;
+
+ if (dormdir)
+ virCgroupRemove(subgroup);
+
+ virCgroupFree(&subgroup);
+ }
+
+ rc = killedAny;
+
+cleanup:
+ virCgroupFree(&subgroup);
+ closedir(dp);
+
+ return rc;
+}
+
+int virCgroupKillRecursive(virCgroupPtr group, int signum)
+{
+ int rc;
+ VIR_DEBUG("group=%p path=%s signum=%d", group, group->path, signum);
+ virHashTablePtr pids = virHashCreateFull(100,
+ NULL,
+ virCgroupPidCode,
+ virCgroupPidEqual,
+ virCgroupPidCopy,
+ NULL);
+
+ rc = virCgroupKillRecursiveInternal(group, signum, pids, false);
+
+ virHashFree(pids);
+
+ return rc;
+}
+
+
+int virCgroupKillPainfully(virCgroupPtr group)
+{
+ int i;
+ int rc;
+ VIR_DEBUG("cgroup=%p path=%s", group, group->path);
+ for (i = 0 ; i < 15 ; i++) {
+ int signum;
+ if (i == 0)
+ signum = SIGTERM;
+ else if (i == 8)
+ signum = SIGKILL;
+ else
+ signum = 0; /* Just check for existence */
+
+ rc = virCgroupKillRecursive(group, signum);
+ VIR_DEBUG("Iteration %d rc=%d", i, rc);
+ /* If rc == -1 we hit error, if 0 we ran out of PIDs */
+ if (rc <= 0)
+ break;
+
+ usleep(200 * 1000);
+ }
+ VIR_DEBUG("Complete %d", rc);
+ return rc;
+}
+
+#else /* !(HAVE_KILL, HAVE_MNTENT_H, HAVE_GETMNTENT_R) */
+int virCgroupKill(virCgroupPtr group ATTRIBUTE_UNUSED,
+ int signum ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+int virCgroupKillRecursive(virCgroupPtr group ATTRIBUTE_UNUSED,
+ int signum ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+
+int virCgroupKillPainfully(virCgroupPtr group ATTRIBUTE_UNUSED)
+{
+ return -ENOSYS;
+}
+#endif /* HAVE_KILL, HAVE_MNTENT_H, HAVE_GETMNTENT_R */
diff --git a/src/util/vircgroup.h b/src/util/vircgroup.h
new file mode 100644
index 0000000..8b6d3b2
--- /dev/null
+++ b/src/util/vircgroup.h
@@ -0,0 +1,167 @@
+/*
+ * vircgroup.h: methods for managing control cgroups
+ *
+ * Copyright (C) 2011-2012 Red Hat, Inc.
+ * Copyright IBM Corp. 2008
+ *
+ * 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/>.
+ *
+ * Authors:
+ * Dan Smith <danms@xxxxxxxxxx>
+ */
+
+#ifndef __VIR_CGROUP_H__
+# define __VIR_CGROUP_H__
+
+struct virCgroup;
+typedef struct virCgroup *virCgroupPtr;
+
+enum {
+ VIR_CGROUP_CONTROLLER_CPU,
+ VIR_CGROUP_CONTROLLER_CPUACCT,
+ VIR_CGROUP_CONTROLLER_CPUSET,
+ VIR_CGROUP_CONTROLLER_MEMORY,
+ VIR_CGROUP_CONTROLLER_DEVICES,
+ VIR_CGROUP_CONTROLLER_FREEZER,
+ VIR_CGROUP_CONTROLLER_BLKIO,
+
+ VIR_CGROUP_CONTROLLER_LAST
+};
+
+VIR_ENUM_DECL(virCgroupController);
+
+int virCgroupForDriver(const char *name,
+ virCgroupPtr *group,
+ bool privileged,
+ bool create);
+
+int virCgroupGetAppRoot(virCgroupPtr *group);
+
+int virCgroupForDomain(virCgroupPtr driver,
+ const char *name,
+ virCgroupPtr *group,
+ bool create);
+
+int virCgroupForVcpu(virCgroupPtr driver,
+ int vcpuid,
+ virCgroupPtr *group,
+ bool create);
+
+int virCgroupForEmulator(virCgroupPtr driver,
+ virCgroupPtr *group,
+ bool create);
+
+int virCgroupPathOfController(virCgroupPtr group,
+ int controller,
+ const char *key,
+ char **path);
+
+int virCgroupAddTask(virCgroupPtr group, pid_t pid);
+
+int virCgroupAddTaskController(virCgroupPtr group,
+ pid_t pid,
+ int controller);
+
+int virCgroupMoveTask(virCgroupPtr src_group,
+ virCgroupPtr dest_group,
+ int controller);
+
+int virCgroupSetBlkioWeight(virCgroupPtr group, unsigned int weight);
+int virCgroupGetBlkioWeight(virCgroupPtr group, unsigned int *weight);
+
+int virCgroupSetBlkioDeviceWeight(virCgroupPtr group,
+ const char *path,
+ unsigned int weight);
+
+int virCgroupSetMemory(virCgroupPtr group, unsigned long long kb);
+int virCgroupGetMemoryUsage(virCgroupPtr group, unsigned long *kb);
+
+int virCgroupSetMemoryHardLimit(virCgroupPtr group, unsigned long long kb);
+int virCgroupGetMemoryHardLimit(virCgroupPtr group, unsigned long long *kb);
+int virCgroupSetMemorySoftLimit(virCgroupPtr group, unsigned long long kb);
+int virCgroupGetMemorySoftLimit(virCgroupPtr group, unsigned long long *kb);
+int virCgroupSetMemSwapHardLimit(virCgroupPtr group, unsigned long long kb);
+int virCgroupGetMemSwapHardLimit(virCgroupPtr group, unsigned long long *kb);
+int virCgroupGetMemSwapUsage(virCgroupPtr group, unsigned long long *kb);
+
+enum {
+ VIR_CGROUP_DEVICE_READ = 1,
+ VIR_CGROUP_DEVICE_WRITE = 2,
+ VIR_CGROUP_DEVICE_MKNOD = 4,
+ VIR_CGROUP_DEVICE_RW = VIR_CGROUP_DEVICE_READ | VIR_CGROUP_DEVICE_WRITE,
+ VIR_CGROUP_DEVICE_RWM = VIR_CGROUP_DEVICE_RW | VIR_CGROUP_DEVICE_MKNOD,
+};
+
+int virCgroupDenyAllDevices(virCgroupPtr group);
+
+int virCgroupAllowDevice(virCgroupPtr group,
+ char type,
+ int major,
+ int minor,
+ int perms);
+int virCgroupAllowDeviceMajor(virCgroupPtr group,
+ char type,
+ int major,
+ int perms);
+int virCgroupAllowDevicePath(virCgroupPtr group,
+ const char *path,
+ int perms);
+
+int virCgroupDenyDevice(virCgroupPtr group,
+ char type,
+ int major,
+ int minor,
+ int perms);
+int virCgroupDenyDeviceMajor(virCgroupPtr group,
+ char type,
+ int major,
+ int perms);
+int virCgroupDenyDevicePath(virCgroupPtr group,
+ const char *path,
+ int perms);
+
+int virCgroupSetCpuShares(virCgroupPtr group, unsigned long long shares);
+int virCgroupGetCpuShares(virCgroupPtr group, unsigned long long *shares);
+
+int virCgroupSetCpuCfsPeriod(virCgroupPtr group, unsigned long long cfs_period);
+int virCgroupGetCpuCfsPeriod(virCgroupPtr group, unsigned long long *cfs_period);
+
+int virCgroupSetCpuCfsQuota(virCgroupPtr group, long long cfs_quota);
+int virCgroupGetCpuCfsQuota(virCgroupPtr group, long long *cfs_quota);
+
+int virCgroupGetCpuacctUsage(virCgroupPtr group, unsigned long long *usage);
+int virCgroupGetCpuacctPercpuUsage(virCgroupPtr group, char **usage);
+int virCgroupGetCpuacctStat(virCgroupPtr group, unsigned long long *user,
+ unsigned long long *sys);
+
+int virCgroupSetFreezerState(virCgroupPtr group, const char *state);
+int virCgroupGetFreezerState(virCgroupPtr group, char **state);
+
+int virCgroupSetCpusetMems(virCgroupPtr group, const char *mems);
+int virCgroupGetCpusetMems(virCgroupPtr group, char **mems);
+
+int virCgroupSetCpusetCpus(virCgroupPtr group, const char *cpus);
+int virCgroupGetCpusetCpus(virCgroupPtr group, char **cpus);
+
+int virCgroupRemove(virCgroupPtr group);
+
+void virCgroupFree(virCgroupPtr *group);
+bool virCgroupMounted(virCgroupPtr cgroup, int controller);
+
+int virCgroupKill(virCgroupPtr group, int signum);
+int virCgroupKillRecursive(virCgroupPtr group, int signum);
+int virCgroupKillPainfully(virCgroupPtr group);
+
+#endif /* __VIR_CGROUP_H__ */
--
1.7.11.7
--
libvir-list mailing list
libvir-list@xxxxxxxxxx
https://www.redhat.com/mailman/listinfo/libvir-list