2012-07-13 11:21:27 +00:00
|
|
|
/*
|
|
|
|
* Copyright (C) 2010-2012 Red Hat, Inc.
|
|
|
|
* Copyright IBM Corp. 2008
|
|
|
|
*
|
|
|
|
* lxc_cgroup.c: LXC cgroup helpers
|
|
|
|
*
|
|
|
|
* 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
|
2012-09-20 22:30:55 +00:00
|
|
|
* License along with this library. If not, see
|
2012-07-21 10:06:23 +00:00
|
|
|
* <http://www.gnu.org/licenses/>.
|
2012-07-13 11:21:27 +00:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <config.h>
|
|
|
|
|
|
|
|
#include "lxc_cgroup.h"
|
|
|
|
#include "lxc_container.h"
|
make /proc/meminfo isolate with host through fuse
with this patch,container's meminfo will be shown based on
containers' mem cgroup.
Right now,it's impossible to virtualize all values in meminfo,
I collect some values such as MemTotal,MemFree,Cached,Active,
Inactive,Active(anon),Inactive(anon),Active(file),Inactive(anon),
Active(file),Inactive(file),Unevictable,SwapTotal,SwapFree.
if I miss something, please let me know.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
2012-11-12 11:52:01 +00:00
|
|
|
#include "virfile.h"
|
2012-12-13 18:21:53 +00:00
|
|
|
#include "virerror.h"
|
2012-12-12 17:59:27 +00:00
|
|
|
#include "virlog.h"
|
2012-12-12 18:06:53 +00:00
|
|
|
#include "viralloc.h"
|
2012-12-03 15:03:47 +00:00
|
|
|
#include "vircgroup.h"
|
2013-04-03 10:36:23 +00:00
|
|
|
#include "virstring.h"
|
2012-07-13 11:21:27 +00:00
|
|
|
|
|
|
|
#define VIR_FROM_THIS VIR_FROM_LXC
|
|
|
|
|
|
|
|
static int virLXCCgroupSetupCpuTune(virDomainDefPtr def,
|
|
|
|
virCgroupPtr cgroup)
|
|
|
|
{
|
|
|
|
int ret = -1;
|
|
|
|
if (def->cputune.shares != 0) {
|
|
|
|
int rc = virCgroupSetCpuShares(cgroup, def->cputune.shares);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set io cpu shares for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (def->cputune.quota != 0) {
|
|
|
|
int rc = virCgroupSetCpuCfsQuota(cgroup, def->cputune.quota);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set io cpu quota for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
if (def->cputune.period != 0) {
|
|
|
|
int rc = virCgroupSetCpuCfsPeriod(cgroup, def->cputune.period);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set io cpu period for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
ret = 0;
|
|
|
|
cleanup:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2013-03-20 03:35:09 +00:00
|
|
|
static int virLXCCgroupSetupCpusetTune(virDomainDefPtr def,
|
|
|
|
virCgroupPtr cgroup,
|
|
|
|
virBitmapPtr nodemask)
|
|
|
|
{
|
|
|
|
int rc = 0;
|
|
|
|
char *mask = NULL;
|
|
|
|
|
|
|
|
if (def->placement_mode != VIR_DOMAIN_CPU_PLACEMENT_MODE_AUTO &&
|
|
|
|
def->cpumask) {
|
|
|
|
mask = virBitmapFormat(def->cpumask);
|
|
|
|
if (!mask) {
|
|
|
|
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
|
|
|
|
_("failed to convert cpumask"));
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = virCgroupSetCpusetCpus(cgroup, mask);
|
|
|
|
if (rc < 0) {
|
|
|
|
virReportSystemError(-rc, "%s",
|
|
|
|
_("Unable to set cpuset.cpus"));
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if ((def->numatune.memory.nodemask ||
|
|
|
|
(def->numatune.memory.placement_mode ==
|
|
|
|
VIR_NUMA_TUNE_MEM_PLACEMENT_MODE_AUTO)) &&
|
|
|
|
def->numatune.memory.mode == VIR_DOMAIN_NUMATUNE_MEM_STRICT) {
|
|
|
|
if (def->numatune.memory.placement_mode ==
|
|
|
|
VIR_NUMA_TUNE_MEM_PLACEMENT_MODE_AUTO)
|
|
|
|
mask = virBitmapFormat(nodemask);
|
|
|
|
else
|
|
|
|
mask = virBitmapFormat(def->numatune.memory.nodemask);
|
|
|
|
|
|
|
|
if (!mask) {
|
|
|
|
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
|
|
|
|
_("failed to convert memory nodemask"));
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = virCgroupSetCpusetMems(cgroup, mask);
|
|
|
|
if (rc < 0)
|
|
|
|
virReportSystemError(-rc, "%s", _("Unable to set cpuset.mems"));
|
|
|
|
}
|
|
|
|
|
|
|
|
cleanup:
|
|
|
|
VIR_FREE(mask);
|
|
|
|
return rc;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
static int virLXCCgroupSetupBlkioTune(virDomainDefPtr def,
|
|
|
|
virCgroupPtr cgroup)
|
|
|
|
{
|
|
|
|
int ret = -1;
|
|
|
|
|
|
|
|
if (def->blkio.weight) {
|
|
|
|
int rc = virCgroupSetBlkioWeight(cgroup, def->blkio.weight);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set Blkio weight for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
cleanup:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int virLXCCgroupSetupMemTune(virDomainDefPtr def,
|
|
|
|
virCgroupPtr cgroup)
|
|
|
|
{
|
|
|
|
int ret = -1;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
rc = virCgroupSetMemory(cgroup, def->mem.max_balloon);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set memory limit for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (def->mem.hard_limit) {
|
|
|
|
rc = virCgroupSetMemoryHardLimit(cgroup, def->mem.hard_limit);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set memory hard limit for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (def->mem.soft_limit) {
|
|
|
|
rc = virCgroupSetMemorySoftLimit(cgroup, def->mem.soft_limit);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set memory soft limit for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
if (def->mem.swap_hard_limit) {
|
|
|
|
rc = virCgroupSetMemSwapHardLimit(cgroup, def->mem.swap_hard_limit);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to set swap hard limit for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
cleanup:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
make /proc/meminfo isolate with host through fuse
with this patch,container's meminfo will be shown based on
containers' mem cgroup.
Right now,it's impossible to virtualize all values in meminfo,
I collect some values such as MemTotal,MemFree,Cached,Active,
Inactive,Active(anon),Inactive(anon),Active(file),Inactive(anon),
Active(file),Inactive(file),Unevictable,SwapTotal,SwapFree.
if I miss something, please let me know.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
2012-11-12 11:52:01 +00:00
|
|
|
static int virLXCCgroupGetMemSwapUsage(virCgroupPtr cgroup,
|
|
|
|
virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
return virCgroupGetMemSwapUsage(cgroup, &meminfo->swapusage);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int virLXCCgroupGetMemSwapTotal(virCgroupPtr cgroup,
|
|
|
|
virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
return virCgroupGetMemSwapHardLimit(cgroup, &meminfo->swaptotal);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int virLXCCgroupGetMemUsage(virCgroupPtr cgroup,
|
|
|
|
virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
unsigned long memUsage;
|
|
|
|
|
|
|
|
ret = virCgroupGetMemoryUsage(cgroup, &memUsage);
|
|
|
|
meminfo->memusage = (unsigned long long) memUsage;
|
|
|
|
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int virLXCCgroupGetMemTotal(virCgroupPtr cgroup,
|
|
|
|
virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
return virCgroupGetMemoryHardLimit(cgroup, &meminfo->memtotal);
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
static int virLXCCgroupGetMemStat(virCgroupPtr cgroup,
|
|
|
|
virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
int ret = 0;
|
|
|
|
FILE *statfd = NULL;
|
|
|
|
char *statFile = NULL;
|
|
|
|
char *line = NULL;
|
|
|
|
size_t n;
|
|
|
|
|
|
|
|
ret = virCgroupPathOfController(cgroup, VIR_CGROUP_CONTROLLER_MEMORY,
|
|
|
|
"memory.stat", &statFile);
|
|
|
|
if (ret != 0) {
|
|
|
|
virReportSystemError(-ret, "%s",
|
|
|
|
_("cannot get the path of MEMORY cgroup controller"));
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
statfd = fopen(statFile, "r");
|
|
|
|
if (statfd == NULL) {
|
|
|
|
ret = -errno;
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
while (getline(&line, &n, statfd) > 0) {
|
|
|
|
|
|
|
|
char *value = strchr(line, ' ');
|
|
|
|
char *nl = value ? strchr(line, '\n') : NULL;
|
|
|
|
unsigned long long stat_value;
|
|
|
|
|
|
|
|
if (!value)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
if (nl)
|
|
|
|
*nl = '\0';
|
|
|
|
|
|
|
|
*value = '\0';
|
|
|
|
|
|
|
|
if (virStrToLong_ull(value + 1, NULL, 10, &stat_value) < 0) {
|
|
|
|
ret = -EINVAL;
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
if (STREQ(line, "cache"))
|
|
|
|
meminfo->cached = stat_value >> 10;
|
|
|
|
else if (STREQ(line, "inactive_anon"))
|
|
|
|
meminfo->inactive_anon = stat_value >> 10;
|
|
|
|
else if (STREQ(line, "active_anon"))
|
|
|
|
meminfo->active_anon = stat_value >> 10;
|
|
|
|
else if (STREQ(line, "inactive_file"))
|
|
|
|
meminfo->inactive_file = stat_value >> 10;
|
|
|
|
else if (STREQ(line, "active_file"))
|
|
|
|
meminfo->active_file = stat_value >> 10;
|
|
|
|
else if (STREQ(line, "unevictable"))
|
|
|
|
meminfo->unevictable = stat_value >> 10;
|
|
|
|
}
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
cleanup:
|
|
|
|
VIR_FREE(line);
|
|
|
|
VIR_FREE(statFile);
|
|
|
|
VIR_FORCE_FCLOSE(statfd);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int virLXCCgroupGetMeminfo(virLXCMeminfoPtr meminfo)
|
|
|
|
{
|
|
|
|
int ret;
|
|
|
|
virCgroupPtr cgroup;
|
|
|
|
|
2013-03-28 16:33:22 +00:00
|
|
|
ret = virCgroupNewSelf(&cgroup);
|
make /proc/meminfo isolate with host through fuse
with this patch,container's meminfo will be shown based on
containers' mem cgroup.
Right now,it's impossible to virtualize all values in meminfo,
I collect some values such as MemTotal,MemFree,Cached,Active,
Inactive,Active(anon),Inactive(anon),Active(file),Inactive(anon),
Active(file),Inactive(file),Unevictable,SwapTotal,SwapFree.
if I miss something, please let me know.
Signed-off-by: Gao feng <gaofeng@cn.fujitsu.com>
2012-11-12 11:52:01 +00:00
|
|
|
if (ret < 0) {
|
|
|
|
virReportSystemError(-ret, "%s",
|
|
|
|
_("Unable to get cgroup for container"));
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = virLXCCgroupGetMemStat(cgroup, meminfo);
|
|
|
|
if (ret < 0) {
|
|
|
|
virReportSystemError(-ret, "%s",
|
|
|
|
_("Unable to get memory cgroup stat info"));
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = virLXCCgroupGetMemTotal(cgroup, meminfo);
|
|
|
|
if (ret < 0) {
|
|
|
|
virReportSystemError(-ret, "%s",
|
|
|
|
_("Unable to get memory cgroup total"));
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = virLXCCgroupGetMemUsage(cgroup, meminfo);
|
|
|
|
if (ret < 0) {
|
|
|
|
virReportSystemError(-ret, "%s",
|
|
|
|
_("Unable to get memory cgroup stat usage"));
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
virLXCCgroupGetMemSwapTotal(cgroup, meminfo);
|
|
|
|
virLXCCgroupGetMemSwapUsage(cgroup, meminfo);
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
cleanup:
|
|
|
|
virCgroupFree(&cgroup);
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
typedef struct _virLXCCgroupDevicePolicy virLXCCgroupDevicePolicy;
|
|
|
|
typedef virLXCCgroupDevicePolicy *virLXCCgroupDevicePolicyPtr;
|
|
|
|
|
|
|
|
struct _virLXCCgroupDevicePolicy {
|
|
|
|
char type;
|
|
|
|
int major;
|
|
|
|
int minor;
|
|
|
|
};
|
|
|
|
|
|
|
|
|
2012-11-23 14:46:18 +00:00
|
|
|
int
|
2013-01-14 22:11:44 +00:00
|
|
|
virLXCSetupHostUsbDeviceCgroup(virUSBDevicePtr dev ATTRIBUTE_UNUSED,
|
2012-11-23 14:46:18 +00:00
|
|
|
const char *path,
|
|
|
|
void *opaque)
|
|
|
|
{
|
|
|
|
virCgroupPtr cgroup = opaque;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
VIR_DEBUG("Process path '%s' for USB device", path);
|
|
|
|
rc = virCgroupAllowDevicePath(cgroup, path,
|
|
|
|
VIR_CGROUP_DEVICE_RW);
|
|
|
|
if (rc < 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to allow device %s"),
|
|
|
|
path);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int
|
2013-01-14 22:11:44 +00:00
|
|
|
virLXCTeardownHostUsbDeviceCgroup(virUSBDevicePtr dev ATTRIBUTE_UNUSED,
|
2012-11-23 14:46:18 +00:00
|
|
|
const char *path,
|
|
|
|
void *opaque)
|
|
|
|
{
|
|
|
|
virCgroupPtr cgroup = opaque;
|
|
|
|
int rc;
|
|
|
|
|
|
|
|
VIR_DEBUG("Process path '%s' for USB device", path);
|
|
|
|
rc = virCgroupDenyDevicePath(cgroup, path,
|
|
|
|
VIR_CGROUP_DEVICE_RW);
|
|
|
|
if (rc < 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to deny device %s"),
|
|
|
|
path);
|
|
|
|
return -1;
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
|
|
|
|
static int virLXCCgroupSetupDeviceACL(virDomainDefPtr def,
|
|
|
|
virCgroupPtr cgroup)
|
|
|
|
{
|
|
|
|
int ret = -1;
|
|
|
|
int rc;
|
|
|
|
size_t i;
|
|
|
|
static virLXCCgroupDevicePolicy devices[] = {
|
|
|
|
{'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_NULL},
|
|
|
|
{'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_ZERO},
|
|
|
|
{'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_FULL},
|
|
|
|
{'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_RANDOM},
|
|
|
|
{'c', LXC_DEV_MAJ_MEMORY, LXC_DEV_MIN_URANDOM},
|
|
|
|
{'c', LXC_DEV_MAJ_TTY, LXC_DEV_MIN_TTY},
|
|
|
|
{'c', LXC_DEV_MAJ_TTY, LXC_DEV_MIN_PTMX},
|
2012-11-14 09:39:04 +00:00
|
|
|
{'c', LXC_DEV_MAJ_FUSE, LXC_DEV_MIN_FUSE},
|
2012-07-13 11:21:27 +00:00
|
|
|
{0, 0, 0}};
|
|
|
|
|
|
|
|
rc = virCgroupDenyAllDevices(cgroup);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to deny devices for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
for (i = 0; devices[i].type != 0; i++) {
|
|
|
|
virLXCCgroupDevicePolicyPtr dev = &devices[i];
|
|
|
|
rc = virCgroupAllowDevice(cgroup,
|
|
|
|
dev->type,
|
|
|
|
dev->major,
|
|
|
|
dev->minor,
|
|
|
|
VIR_CGROUP_DEVICE_RWM);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to allow device %c:%d:%d for domain %s"),
|
|
|
|
dev->type, dev->major, dev->minor, def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-05-21 08:03:33 +00:00
|
|
|
for (i = 0; i < def->ndisks; i++) {
|
2012-11-22 14:33:48 +00:00
|
|
|
if (def->disks[i]->type != VIR_DOMAIN_DISK_TYPE_BLOCK)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
rc = virCgroupAllowDevicePath(cgroup,
|
|
|
|
def->disks[i]->src,
|
|
|
|
(def->disks[i]->readonly ?
|
|
|
|
VIR_CGROUP_DEVICE_READ :
|
|
|
|
VIR_CGROUP_DEVICE_RW) |
|
|
|
|
VIR_CGROUP_DEVICE_MKNOD);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to allow device %s for domain %s"),
|
|
|
|
def->disks[i]->src, def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2013-05-21 08:03:33 +00:00
|
|
|
for (i = 0; i < def->nfss; i++) {
|
2012-07-13 11:21:27 +00:00
|
|
|
if (def->fss[i]->type != VIR_DOMAIN_FS_TYPE_BLOCK)
|
|
|
|
continue;
|
|
|
|
|
|
|
|
rc = virCgroupAllowDevicePath(cgroup,
|
|
|
|
def->fss[i]->src,
|
|
|
|
def->fss[i]->readonly ?
|
|
|
|
VIR_CGROUP_DEVICE_READ :
|
|
|
|
VIR_CGROUP_DEVICE_RW);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to allow device %s for domain %s"),
|
|
|
|
def->fss[i]->src, def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2012-11-23 14:46:18 +00:00
|
|
|
for (i = 0; i < def->nhostdevs; i++) {
|
|
|
|
virDomainHostdevDefPtr hostdev = def->hostdevs[i];
|
2013-01-14 22:11:44 +00:00
|
|
|
virUSBDevicePtr usb;
|
2012-11-23 14:46:18 +00:00
|
|
|
|
2012-11-28 16:13:27 +00:00
|
|
|
switch (hostdev->mode) {
|
|
|
|
case VIR_DOMAIN_HOSTDEV_MODE_SUBSYS:
|
|
|
|
if (hostdev->source.subsys.type != VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB)
|
|
|
|
continue;
|
|
|
|
if (hostdev->missing)
|
|
|
|
continue;
|
|
|
|
|
2013-01-14 22:11:44 +00:00
|
|
|
if ((usb = virUSBDeviceNew(hostdev->source.subsys.u.usb.bus,
|
|
|
|
hostdev->source.subsys.u.usb.device,
|
|
|
|
NULL)) == NULL)
|
2012-11-28 16:13:27 +00:00
|
|
|
goto cleanup;
|
|
|
|
|
2013-01-14 22:11:44 +00:00
|
|
|
if (virUSBDeviceFileIterate(usb, virLXCSetupHostUsbDeviceCgroup,
|
|
|
|
cgroup) < 0) {
|
|
|
|
virUSBDeviceFree(usb);
|
2012-11-28 16:13:27 +00:00
|
|
|
goto cleanup;
|
2013-02-05 15:14:46 +00:00
|
|
|
}
|
2013-01-14 22:11:44 +00:00
|
|
|
virUSBDeviceFree(usb);
|
2012-11-28 16:13:27 +00:00
|
|
|
break;
|
|
|
|
case VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES:
|
|
|
|
switch (hostdev->source.caps.type) {
|
|
|
|
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_STORAGE:
|
|
|
|
if (virCgroupAllowDevicePath(cgroup,
|
|
|
|
hostdev->source.caps.u.storage.block,
|
|
|
|
VIR_CGROUP_DEVICE_RW |
|
|
|
|
VIR_CGROUP_DEVICE_MKNOD) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
break;
|
2012-11-28 18:07:47 +00:00
|
|
|
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_MISC:
|
|
|
|
if (virCgroupAllowDevicePath(cgroup,
|
|
|
|
hostdev->source.caps.u.misc.chardev,
|
|
|
|
VIR_CGROUP_DEVICE_RW |
|
|
|
|
VIR_CGROUP_DEVICE_MKNOD) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
break;
|
2012-11-28 16:13:27 +00:00
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
|
|
|
default:
|
|
|
|
break;
|
|
|
|
}
|
2012-11-23 14:46:18 +00:00
|
|
|
}
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
rc = virCgroupAllowDeviceMajor(cgroup, 'c', LXC_DEV_MAJ_PTY,
|
|
|
|
VIR_CGROUP_DEVICE_RWM);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to allow PTY devices for domain %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
cleanup:
|
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
virCgroupPtr virLXCCgroupCreate(virDomainDefPtr def, bool startup)
|
2012-07-13 11:21:27 +00:00
|
|
|
{
|
2012-11-23 10:42:18 +00:00
|
|
|
int rc;
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
virCgroupPtr parent = NULL;
|
|
|
|
virCgroupPtr cgroup = NULL;
|
2012-07-13 11:21:27 +00:00
|
|
|
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
if (!def->resource && startup) {
|
|
|
|
virDomainResourceDefPtr res;
|
|
|
|
|
|
|
|
if (VIR_ALLOC(res) < 0) {
|
|
|
|
virReportOOMError();
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
2013-05-03 12:43:39 +00:00
|
|
|
if (VIR_STRDUP(res->partition, "/machine") < 0) {
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
VIR_FREE(res);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
def->resource = res;
|
2012-07-13 11:21:27 +00:00
|
|
|
}
|
|
|
|
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
if (def->resource &&
|
|
|
|
def->resource->partition) {
|
|
|
|
if (def->resource->partition[0] != '/') {
|
|
|
|
virReportError(VIR_ERR_CONFIG_UNSUPPORTED,
|
|
|
|
_("Resource partition '%s' must start with '/'"),
|
|
|
|
def->resource->partition);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
/* We only auto-create the default partition. In other
|
|
|
|
* cases we expec the sysadmin/app to have done so */
|
|
|
|
rc = virCgroupNewPartition(def->resource->partition,
|
2013-04-22 16:11:11 +00:00
|
|
|
STREQ(def->resource->partition, "/machine"),
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
-1,
|
|
|
|
&parent);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to initialize %s cgroup"),
|
|
|
|
def->resource->partition);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = virCgroupNewDomainPartition(parent,
|
|
|
|
"lxc",
|
|
|
|
def->name,
|
|
|
|
true,
|
|
|
|
&cgroup);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to create cgroup for %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
rc = virCgroupNewDriver("lxc",
|
|
|
|
true,
|
|
|
|
-1,
|
|
|
|
&parent);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to create cgroup for %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
rc = virCgroupNewDomainDriver(parent,
|
|
|
|
def->name,
|
|
|
|
true,
|
|
|
|
&cgroup);
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to create cgroup for %s"),
|
|
|
|
def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
2012-07-13 11:21:27 +00:00
|
|
|
}
|
|
|
|
|
2013-03-21 14:40:29 +00:00
|
|
|
cleanup:
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
virCgroupFree(&parent);
|
2013-03-21 14:40:29 +00:00
|
|
|
return cgroup;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
virCgroupPtr virLXCCgroupJoin(virDomainDefPtr def)
|
|
|
|
{
|
|
|
|
virCgroupPtr cgroup = NULL;
|
|
|
|
int ret = -1;
|
|
|
|
int rc;
|
|
|
|
|
Change default cgroup layout for QEMU/LXC and honour XML config
Historically QEMU/LXC guests have been placed in a cgroup layout
that is
$LOCATION-OF-LIBVIRTD/libvirt/{qemu,lxc}/$VMNAME
This is bad for a number of reasons
- The cgroup hierarchy gets very deep which seriously
impacts kernel performance due to cgroups scalability
limitations.
- It is hard to setup cgroup policies which apply across
services and virtual machines, since all VMs are underneath
the libvirtd service.
To address this the default cgroup location is changed to
be
/system/$VMNAME.{lxc,qemu}.libvirt
This puts virtual machines at the same level in the hierarchy
as system services, allowing consistent policy to be setup
across all of them.
This also honours the new resource partition location from the
XML configuration, for example
<resource>
<partition>/virtualmachines/production</partitions>
</resource>
will result in the VM being placed at
/virtualmachines/production/$VMNAME.{lxc,qemu}.libvirt
NB, with the exception of the default, /system, path which
is intended to always exist, libvirt will not attempt to
auto-create the partitions in the XML. It is the responsibility
of the admin/app to configure the partitions. Later libvirt
APIs will provide a way todo this.
Signed-off-by: Daniel P. Berrange <berrange@redhat.com>
2013-04-03 10:01:49 +00:00
|
|
|
if (!(cgroup = virLXCCgroupCreate(def, true)))
|
2013-03-21 14:40:29 +00:00
|
|
|
return NULL;
|
|
|
|
|
2013-03-14 12:47:28 +00:00
|
|
|
rc = virCgroupAddTask(cgroup, getpid());
|
|
|
|
if (rc != 0) {
|
|
|
|
virReportSystemError(-rc,
|
|
|
|
_("Unable to add task %d to cgroup for domain %s"),
|
|
|
|
getpid(), def->name);
|
|
|
|
goto cleanup;
|
|
|
|
}
|
|
|
|
|
|
|
|
ret = 0;
|
|
|
|
|
|
|
|
cleanup:
|
|
|
|
if (ret < 0) {
|
|
|
|
virCgroupFree(&cgroup);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
return cgroup;
|
|
|
|
}
|
|
|
|
|
|
|
|
|
|
|
|
int virLXCCgroupSetup(virDomainDefPtr def,
|
2013-03-20 03:35:09 +00:00
|
|
|
virCgroupPtr cgroup,
|
|
|
|
virBitmapPtr nodemask)
|
2013-03-14 12:47:28 +00:00
|
|
|
{
|
|
|
|
int ret = -1;
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
if (virLXCCgroupSetupCpuTune(def, cgroup) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
|
2013-03-20 03:35:09 +00:00
|
|
|
if (virLXCCgroupSetupCpusetTune(def, cgroup, nodemask) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
if (virLXCCgroupSetupBlkioTune(def, cgroup) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
|
|
|
|
if (virLXCCgroupSetupMemTune(def, cgroup) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
|
|
|
|
if (virLXCCgroupSetupDeviceACL(def, cgroup) < 0)
|
|
|
|
goto cleanup;
|
|
|
|
|
2012-11-23 10:42:18 +00:00
|
|
|
ret = 0;
|
|
|
|
|
2012-07-13 11:21:27 +00:00
|
|
|
cleanup:
|
2012-11-23 10:42:18 +00:00
|
|
|
return ret;
|
2012-07-13 11:21:27 +00:00
|
|
|
}
|