libvirt/src/security/security_selinux.c

3090 lines
93 KiB
C
Raw Normal View History

/*
* Copyright (C) 2008-2014 Red Hat, Inc.
*
* 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:
* James Morris <jmorris@namei.org>
* Dan Walsh <dwalsh@redhat.com>
*
* SELinux security driver.
*/
#include <config.h>
#include <selinux/selinux.h>
#include <selinux/context.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <fcntl.h>
#if HAVE_SELINUX_LABEL_H
# include <selinux/label.h>
#endif
#include "security_driver.h"
#include "security_selinux.h"
#include "virerror.h"
2012-12-12 18:06:53 +00:00
#include "viralloc.h"
2012-12-12 17:59:27 +00:00
#include "virlog.h"
#include "virmdev.h"
2012-12-13 14:52:25 +00:00
#include "virpci.h"
2012-12-12 17:04:51 +00:00
#include "virusb.h"
#include "virscsi.h"
#include "virscsivhost.h"
#include "virstoragefile.h"
#include "virfile.h"
#include "virhash.h"
#include "virrandom.h"
2012-12-12 16:35:35 +00:00
#include "virconf.h"
#include "virtpm.h"
#include "virstring.h"
2009-03-03 15:18:24 +00:00
#define VIR_FROM_THIS VIR_FROM_SECURITY
VIR_LOG_INIT("security.security_selinux");
#define MAX_CONTEXT 1024
typedef struct _virSecuritySELinuxData virSecuritySELinuxData;
typedef virSecuritySELinuxData *virSecuritySELinuxDataPtr;
struct _virSecuritySELinuxData {
char *domain_context;
char *alt_domain_context;
char *file_context;
char *content_context;
virHashTablePtr mcs;
bool skipAllLabel;
#if HAVE_SELINUX_LABEL_H
struct selabel_handle *label_handle;
#endif
};
/* Data structure to pass to various callbacks so we have everything we need */
typedef struct _virSecuritySELinuxCallbackData virSecuritySELinuxCallbackData;
typedef virSecuritySELinuxCallbackData *virSecuritySELinuxCallbackDataPtr;
struct _virSecuritySELinuxCallbackData {
virSecurityManagerPtr mgr;
virDomainDefPtr def;
};
typedef struct _virSecuritySELinuxContextItem virSecuritySELinuxContextItem;
typedef virSecuritySELinuxContextItem *virSecuritySELinuxContextItemPtr;
struct _virSecuritySELinuxContextItem {
char *path;
char *tcon;
bool optional;
};
typedef struct _virSecuritySELinuxContextList virSecuritySELinuxContextList;
typedef virSecuritySELinuxContextList *virSecuritySELinuxContextListPtr;
struct _virSecuritySELinuxContextList {
bool privileged;
virSecuritySELinuxContextItemPtr *items;
size_t nItems;
};
#define SECURITY_SELINUX_VOID_DOI "0"
#define SECURITY_SELINUX_NAME "selinux"
static int
virSecuritySELinuxRestoreTPMFileLabelInt(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainTPMDefPtr tpm);
virThreadLocal contextList;
static void
virSecuritySELinuxContextItemFree(virSecuritySELinuxContextItemPtr item)
{
if (!item)
return;
VIR_FREE(item->path);
VIR_FREE(item->tcon);
VIR_FREE(item);
}
static int
virSecuritySELinuxContextListAppend(virSecuritySELinuxContextListPtr list,
const char *path,
const char *tcon,
bool optional)
{
int ret = -1;
virSecuritySELinuxContextItemPtr item = NULL;
if (VIR_ALLOC(item) < 0)
return -1;
if (VIR_STRDUP(item->path, path) < 0 || VIR_STRDUP(item->tcon, tcon) < 0)
goto cleanup;
item->optional = optional;
if (VIR_APPEND_ELEMENT(list->items, list->nItems, item) < 0)
goto cleanup;
ret = 0;
cleanup:
virSecuritySELinuxContextItemFree(item);
return ret;
}
static void
virSecuritySELinuxContextListFree(void *opaque)
{
virSecuritySELinuxContextListPtr list = opaque;
size_t i;
if (!list)
return;
for (i = 0; i < list->nItems; i++)
virSecuritySELinuxContextItemFree(list->items[i]);
VIR_FREE(list);
}
/**
* virSecuritySELinuxTransactionAppend:
* @path: Path to chown
* @tcon: target context
* @optional: true if setting @tcon is optional
*
* Appends an entry onto transaction list.
*
* Returns: 1 in case of successful append
* 0 if there is no transaction enabled
* -1 otherwise.
*/
static int
virSecuritySELinuxTransactionAppend(const char *path,
const char *tcon,
bool optional)
{
virSecuritySELinuxContextListPtr list;
list = virThreadLocalGet(&contextList);
if (!list)
return 0;
if (virSecuritySELinuxContextListAppend(list, path, tcon, optional) < 0)
return -1;
return 1;
}
static int virSecuritySELinuxSetFileconHelper(const char *path,
const char *tcon,
bool optional,
bool privileged);
/**
* virSecuritySELinuxTransactionRun:
* @pid: process pid
* @opaque: opaque data
*
* This is the callback that runs in the same namespace as the domain we are
* relabelling. For given transaction (@opaque) it relabels all the paths on
* the list.
*
* Returns: 0 on success
* -1 otherwise.
*/
static int
virSecuritySELinuxTransactionRun(pid_t pid ATTRIBUTE_UNUSED,
void *opaque)
{
virSecuritySELinuxContextListPtr list = opaque;
size_t i;
for (i = 0; i < list->nItems; i++) {
virSecuritySELinuxContextItemPtr item = list->items[i];
/* TODO Implement rollback */
if (virSecuritySELinuxSetFileconHelper(item->path,
item->tcon,
item->optional,
list->privileged) < 0)
return -1;
}
return 0;
}
/*
* Returns 0 on success, 1 if already reserved, or -1 on fatal error
*/
static int
virSecuritySELinuxMCSAdd(virSecurityManagerPtr mgr,
const char *mcs)
{
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
if (virHashLookup(data->mcs, mcs))
return 1;
if (virHashAddEntry(data->mcs, mcs, (void*)0x1) < 0)
return -1;
return 0;
}
static void
virSecuritySELinuxMCSRemove(virSecurityManagerPtr mgr,
const char *mcs)
{
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
virHashRemoveEntry(data->mcs, mcs);
}
static char *
virSecuritySELinuxMCSFind(virSecurityManagerPtr mgr,
const char *sens,
int catMin,
int catMax)
{
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
int catRange;
char *mcs = NULL;
/* +1 since virRandomInt range is exclusive of the upper bound */
catRange = (catMax - catMin) + 1;
if (catRange < 8) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Category range c%d-c%d too small"),
catMin, catMax);
return NULL;
}
VIR_DEBUG("Using sensitivity level '%s' cat min %d max %d range %d",
sens, catMin, catMax, catRange);
for (;;) {
int c1 = virRandomInt(catRange);
int c2 = virRandomInt(catRange);
VIR_DEBUG("Try cat %s:c%d,c%d", sens, c1 + catMin, c2 + catMin);
if (c1 == c2) {
if (virAsprintf(&mcs, "%s:c%d", sens, catMin + c1) < 0)
return NULL;
} else {
if (c1 > c2) {
int t = c1;
c1 = c2;
c2 = t;
}
if (virAsprintf(&mcs, "%s:c%d,c%d", sens, catMin + c1, catMin + c2) < 0)
return NULL;
}
if (virHashLookup(data->mcs, mcs) == NULL)
break;
VIR_FREE(mcs);
}
return mcs;
}
/*
* This needs to cope with several styles of range
*
* system_u:system_r:virtd_t
* system_u:system_r:virtd_t:s0
* system_u:system_r:virtd_t:s0-s0
* system_u:system_r:virtd_t:s0-s0:c0.c1023
*
* In the first case we'll assume s0:c0.c1023 and
* in the next two cases, we'll assume c0.c1023 for
* the category part, since that's what we're really
* interested in. This won't work in Enforcing mode,
* but will prevent libvirtd breaking in Permissive
* mode when run with a weird process label.
*/
static int
virSecuritySELinuxMCSGetProcessRange(char **sens,
int *catMin,
int *catMax)
{
security_context_t ourSecContext = NULL;
context_t ourContext = NULL;
char *cat = NULL;
char *tmp;
const char *contextRange;
int ret = -1;
if (getcon_raw(&ourSecContext) < 0) {
virReportSystemError(errno, "%s",
_("Unable to get current process SELinux context"));
goto cleanup;
}
if (!(ourContext = context_new(ourSecContext))) {
virReportSystemError(errno,
_("Unable to parse current SELinux context '%s'"),
ourSecContext);
goto cleanup;
}
if (!(contextRange = context_range_get(ourContext)))
contextRange = "s0";
if (VIR_STRDUP(*sens, contextRange) < 0)
goto cleanup;
/* Find and blank out the category part (if any) */
tmp = strchr(*sens, ':');
if (tmp) {
*tmp = '\0';
cat = tmp + 1;
}
/* Find and blank out the sensitivity upper bound */
if ((tmp = strchr(*sens, '-')))
*tmp = '\0';
/* sens now just contains the sensitivity lower bound */
/* If there was no category part, just assume c0.c1023 */
if (!cat) {
*catMin = 0;
*catMax = 1023;
ret = 0;
goto cleanup;
}
/* Find & extract category min */
tmp = cat;
if (tmp[0] != 'c') {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Cannot parse category in %s"),
cat);
goto cleanup;
}
tmp++;
if (virStrToLong_i(tmp, &tmp, 10, catMin) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Cannot parse category in %s"),
cat);
goto cleanup;
}
/* We *must* have a pair of categories otherwise
* there's no range to allocate VM categories from */
if (!tmp[0]) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("No category range available"));
goto cleanup;
}
/* Find & extract category max (if any) */
if (tmp[0] != '.') {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Cannot parse category in %s"),
cat);
goto cleanup;
}
tmp++;
if (tmp[0] != 'c') {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Cannot parse category in %s"),
cat);
goto cleanup;
}
tmp++;
if (virStrToLong_i(tmp, &tmp, 10, catMax) < 0) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("Cannot parse category in %s"),
cat);
goto cleanup;
}
ret = 0;
cleanup:
if (ret < 0)
VIR_FREE(*sens);
freecon(ourSecContext);
context_free(ourContext);
return ret;
}
static char *
virSecuritySELinuxContextAddRange(security_context_t src,
security_context_t dst)
{
char *str = NULL;
char *ret = NULL;
context_t srccon = NULL;
context_t dstcon = NULL;
if (!src || !dst)
return ret;
if (!(srccon = context_new(src)) || !(dstcon = context_new(dst))) {
virReportSystemError(errno, "%s",
_("unable to allocate security context"));
goto cleanup;
}
if (context_range_set(dstcon, context_range_get(srccon)) == -1) {
virReportSystemError(errno,
_("unable to set security context range '%s'"), dst);
goto cleanup;
}
if (!(str = context_str(dstcon))) {
virReportSystemError(errno, "%s",
_("Unable to format SELinux context"));
goto cleanup;
}
ignore_value(VIR_STRDUP(ret, str));
cleanup:
if (srccon) context_free(srccon);
if (dstcon) context_free(dstcon);
return ret;
}
static char *
virSecuritySELinuxGenNewContext(const char *basecontext,
const char *mcs,
bool isObjectContext)
{
context_t context = NULL;
char *ret = NULL;
char *str;
security_context_t ourSecContext = NULL;
context_t ourContext = NULL;
VIR_DEBUG("basecontext=%s mcs=%s isObjectContext=%d",
basecontext, mcs, isObjectContext);
if (getcon_raw(&ourSecContext) < 0) {
virReportSystemError(errno, "%s",
_("Unable to get current process SELinux context"));
goto cleanup;
}
if (!(ourContext = context_new(ourSecContext))) {
virReportSystemError(errno,
_("Unable to parse current SELinux context '%s'"),
ourSecContext);
goto cleanup;
}
VIR_DEBUG("process=%s", ourSecContext);
if (!(context = context_new(basecontext))) {
virReportSystemError(errno,
_("Unable to parse base SELinux context '%s'"),
basecontext);
goto cleanup;
}
if (context_user_set(context,
context_user_get(ourContext)) != 0) {
virReportSystemError(errno,
_("Unable to set SELinux context user '%s'"),
context_user_get(ourContext));
goto cleanup;
}
if (!isObjectContext &&
context_role_set(context,
context_role_get(ourContext)) != 0) {
virReportSystemError(errno,
_("Unable to set SELinux context role '%s'"),
context_role_get(ourContext));
goto cleanup;
}
if (context_range_set(context, mcs) != 0) {
virReportSystemError(errno,
_("Unable to set SELinux context MCS '%s'"),
mcs);
goto cleanup;
}
if (!(str = context_str(context))) {
virReportSystemError(errno, "%s",
_("Unable to format SELinux context"));
goto cleanup;
}
if (VIR_STRDUP(ret, str) < 0)
goto cleanup;
VIR_DEBUG("Generated context '%s'", ret);
cleanup:
freecon(ourSecContext);
context_free(ourContext);
context_free(context);
return ret;
}
#ifdef HAVE_SELINUX_LXC_CONTEXTS_PATH
static int
virSecuritySELinuxLXCInitialize(virSecurityManagerPtr mgr)
{
virConfPtr selinux_conf;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
data->skipAllLabel = true;
# if HAVE_SELINUX_LABEL_H
data->label_handle = selabel_open(SELABEL_CTX_FILE, NULL, 0);
if (!data->label_handle) {
virReportSystemError(errno, "%s",
_("cannot open SELinux label_handle"));
return -1;
}
# endif
if (!(selinux_conf = virConfReadFile(selinux_lxc_contexts_path(), 0)))
goto error;
if (virConfGetValueString(selinux_conf, "process", &data->domain_context) < 0)
goto error;
if (!data->domain_context) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("missing 'process' value in selinux lxc contexts file '%s'"),
selinux_lxc_contexts_path());
goto error;
}
if (virConfGetValueString(selinux_conf, "file", &data->file_context) < 0)
goto error;
if (!data->file_context) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("missing 'file' value in selinux lxc contexts file '%s'"),
selinux_lxc_contexts_path());
goto error;
}
if (virConfGetValueString(selinux_conf, "content", &data->content_context) < 0)
goto error;
if (!data->content_context) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("missing 'content' value in selinux lxc contexts file '%s'"),
selinux_lxc_contexts_path());
goto error;
}
if (!(data->mcs = virHashCreate(10, NULL)))
goto error;
virConfFree(selinux_conf);
return 0;
error:
# if HAVE_SELINUX_LABEL_H
selabel_close(data->label_handle);
data->label_handle = NULL;
# endif
virConfFree(selinux_conf);
VIR_FREE(data->domain_context);
VIR_FREE(data->file_context);
VIR_FREE(data->content_context);
virHashFree(data->mcs);
return -1;
}
#else
static int
virSecuritySELinuxLXCInitialize(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED)
{
virReportSystemError(ENOSYS, "%s",
_("libselinux does not support LXC contexts path"));
return -1;
}
#endif
static int
virSecuritySELinuxQEMUInitialize(virSecurityManagerPtr mgr)
{
char *ptr;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
data->skipAllLabel = false;
#if HAVE_SELINUX_LABEL_H
data->label_handle = selabel_open(SELABEL_CTX_FILE, NULL, 0);
if (!data->label_handle) {
virReportSystemError(errno, "%s",
_("cannot open SELinux label_handle"));
return -1;
}
#endif
if (virFileReadAll(selinux_virtual_domain_context_path(), MAX_CONTEXT, &(data->domain_context)) < 0) {
virReportSystemError(errno,
_("cannot read SELinux virtual domain context file '%s'"),
selinux_virtual_domain_context_path());
goto error;
}
ptr = strchrnul(data->domain_context, '\n');
if (ptr && *ptr == '\n') {
*ptr = '\0';
ptr++;
if (*ptr != '\0') {
if (VIR_STRDUP(data->alt_domain_context, ptr) < 0)
goto error;
ptr = strchrnul(data->alt_domain_context, '\n');
if (ptr && *ptr == '\n')
*ptr = '\0';
}
}
VIR_DEBUG("Loaded domain context '%s', alt domain context '%s'",
data->domain_context, NULLSTR(data->alt_domain_context));
if (virFileReadAll(selinux_virtual_image_context_path(), 2*MAX_CONTEXT, &(data->file_context)) < 0) {
virReportSystemError(errno,
_("cannot read SELinux virtual image context file %s"),
selinux_virtual_image_context_path());
goto error;
}
ptr = strchrnul(data->file_context, '\n');
if (ptr && *ptr == '\n') {
*ptr = '\0';
if (VIR_STRDUP(data->content_context, ptr + 1) < 0)
goto error;
ptr = strchrnul(data->content_context, '\n');
if (ptr && *ptr == '\n')
*ptr = '\0';
}
VIR_DEBUG("Loaded file context '%s', content context '%s'",
data->file_context, data->content_context);
if (!(data->mcs = virHashCreate(10, NULL)))
goto error;
return 0;
error:
#if HAVE_SELINUX_LABEL_H
selabel_close(data->label_handle);
data->label_handle = NULL;
#endif
VIR_FREE(data->domain_context);
VIR_FREE(data->alt_domain_context);
VIR_FREE(data->file_context);
VIR_FREE(data->content_context);
virHashFree(data->mcs);
return -1;
}
static int
virSecuritySELinuxInitialize(virSecurityManagerPtr mgr)
{
VIR_DEBUG("SELinuxInitialize %s", virSecurityManagerGetDriver(mgr));
if (virThreadLocalInit(&contextList,
virSecuritySELinuxContextListFree) < 0) {
virReportSystemError(errno, "%s",
_("Unable to initialize thread local variable"));
return -1;
}
if (STREQ(virSecurityManagerGetDriver(mgr), "LXC")) {
return virSecuritySELinuxLXCInitialize(mgr);
} else {
return virSecuritySELinuxQEMUInitialize(mgr);
}
}
static int
virSecuritySELinuxGenLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def)
{
int rc = -1;
char *mcs = NULL;
char *scontext = NULL;
context_t ctx = NULL;
const char *range;
virSecurityLabelDefPtr seclabel;
virSecuritySELinuxDataPtr data;
const char *baselabel;
char *sens = NULL;
int catMin, catMax;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
2013-07-02 16:27:09 +00:00
if (seclabel == NULL)
return 0;
data = virSecurityManagerGetPrivateData(mgr);
VIR_DEBUG("label=%s", virSecurityManagerGetDriver(mgr));
if (seclabel->type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
seclabel->label) {
2013-07-02 16:27:09 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("security label already defined for VM"));
return rc;
2009-03-03 15:18:24 +00:00
}
if (seclabel->imagelabel) {
2013-07-02 16:27:09 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("security image label already defined for VM"));
return rc;
}
if (seclabel->model &&
STRNEQ(seclabel->model, SECURITY_SELINUX_NAME)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label model %s is not supported with selinux"),
seclabel->model);
return rc;
}
VIR_DEBUG("type=%d", seclabel->type);
switch (seclabel->type) {
case VIR_DOMAIN_SECLABEL_STATIC:
if (!(ctx = context_new(seclabel->label))) {
virReportSystemError(errno,
_("unable to allocate socket security context '%s'"),
seclabel->label);
return rc;
}
2013-07-02 16:27:09 +00:00
if (!(range = context_range_get(ctx))) {
virReportSystemError(errno, "%s", _("unable to get selinux context range"));
goto cleanup;
}
if (VIR_STRDUP(mcs, range) < 0)
goto cleanup;
break;
case VIR_DOMAIN_SECLABEL_DYNAMIC:
if (virSecuritySELinuxMCSGetProcessRange(&sens,
&catMin,
&catMax) < 0)
goto cleanup;
if (!(mcs = virSecuritySELinuxMCSFind(mgr,
sens,
catMin,
catMax)))
goto cleanup;
if (virSecuritySELinuxMCSAdd(mgr, mcs) < 0)
goto cleanup;
baselabel = seclabel->baselabel;
if (!baselabel) {
if (def->virtType == VIR_DOMAIN_VIRT_QEMU) {
if (data->alt_domain_context == NULL) {
static bool warned;
if (!warned) {
VIR_WARN("SELinux policy does not define a domain type for QEMU TCG. "
"Guest startup may be denied due to missing 'execmem' privilege "
"unless the 'virt_use_execmem' policy boolean is enabled");
warned = true;
}
baselabel = data->domain_context;
} else {
baselabel = data->alt_domain_context;
}
} else {
baselabel = data->domain_context;
}
}
seclabel->label = virSecuritySELinuxGenNewContext(baselabel, mcs, false);
if (!seclabel->label)
goto cleanup;
break;
case VIR_DOMAIN_SECLABEL_NONE:
if (virSecuritySELinuxMCSGetProcessRange(&sens,
&catMin,
&catMax) < 0)
goto cleanup;
if (VIR_STRDUP(mcs, sens) < 0)
goto cleanup;
break;
default:
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unexpected security label type '%s'"),
virDomainSeclabelTypeToString(seclabel->type));
goto cleanup;
2009-03-03 15:18:24 +00:00
}
/* always generate a image label, needed to label new objects */
seclabel->imagelabel = virSecuritySELinuxGenNewContext(data->file_context,
mcs,
true);
if (!seclabel->imagelabel)
goto cleanup;
if (!seclabel->model &&
VIR_STRDUP(seclabel->model, SECURITY_SELINUX_NAME) < 0)
goto cleanup;
2009-03-03 15:18:24 +00:00
rc = 0;
cleanup:
if (rc != 0) {
if (seclabel->type == VIR_DOMAIN_SECLABEL_DYNAMIC)
VIR_FREE(seclabel->label);
VIR_FREE(seclabel->imagelabel);
if (seclabel->type == VIR_DOMAIN_SECLABEL_DYNAMIC &&
!seclabel->baselabel)
VIR_FREE(seclabel->model);
}
if (ctx)
context_free(ctx);
2009-03-03 15:18:24 +00:00
VIR_FREE(scontext);
VIR_FREE(mcs);
VIR_FREE(sens);
VIR_DEBUG("model=%s label=%s imagelabel=%s baselabel=%s",
NULLSTR(seclabel->model),
NULLSTR(seclabel->label),
NULLSTR(seclabel->imagelabel),
NULLSTR(seclabel->baselabel));
return rc;
}
static int
virSecuritySELinuxReserveLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
pid_t pid)
{
security_context_t pctx;
context_t ctx = NULL;
const char *mcs;
int rv;
virSecurityLabelDefPtr seclabel;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel ||
seclabel->type == VIR_DOMAIN_SECLABEL_NONE ||
seclabel->type == VIR_DOMAIN_SECLABEL_STATIC)
return 0;
if (getpidcon_raw(pid, &pctx) == -1) {
virReportSystemError(errno,
_("unable to get PID %d security context"), pid);
return -1;
}
ctx = context_new(pctx);
freecon(pctx);
if (!ctx)
goto error;
mcs = context_range_get(ctx);
if (!mcs)
goto error;
if ((rv = virSecuritySELinuxMCSAdd(mgr, mcs)) < 0)
goto error;
if (rv == 1) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("MCS level for existing domain label %s already reserved"),
(char*)pctx);
goto error;
}
context_free(ctx);
return 0;
error:
context_free(ctx);
return -1;
}
static int
virSecuritySELinuxDriverProbe(const char *virtDriver)
{
if (is_selinux_enabled() <= 0)
return SECURITY_DRIVER_DISABLE;
if (virtDriver && STREQ(virtDriver, "LXC")) {
#if HAVE_SELINUX_LXC_CONTEXTS_PATH
if (!virFileExists(selinux_lxc_contexts_path()))
#endif
return SECURITY_DRIVER_DISABLE;
}
return SECURITY_DRIVER_ENABLE;
}
static int
virSecuritySELinuxDriverOpen(virSecurityManagerPtr mgr)
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
{
return virSecuritySELinuxInitialize(mgr);
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
}
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
static int
virSecuritySELinuxDriverClose(virSecurityManagerPtr mgr)
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
{
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
if (!data)
return 0;
#if HAVE_SELINUX_LABEL_H
if (data->label_handle)
selabel_close(data->label_handle);
#endif
virHashFree(data->mcs);
VIR_FREE(data->domain_context);
VIR_FREE(data->alt_domain_context);
VIR_FREE(data->file_context);
VIR_FREE(data->content_context);
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
return 0;
}
static const char *
virSecuritySELinuxGetModel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED)
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
{
return SECURITY_SELINUX_NAME;
}
static const char *
virSecuritySELinuxGetDOI(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED)
{
/*
* Where will the DOI come from? SELinux configuration, or qemu
* configuration? For the moment, we'll just set it to "0".
*/
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
return SECURITY_SELINUX_VOID_DOI;
}
/**
* virSecuritySELinuxTransactionStart:
* @mgr: security manager
*
* Starts a new transaction. In transaction nothing is changed context
* until TransactionCommit() is called. This is implemented as a list
* that is appended to whenever setfilecon() would be called. Since
* secdriver APIs can be called from multiple threads (to work over
* different domains) the pointer to the list is stored in thread local
* variable.
*
* Returns 0 on success,
* -1 otherwise.
*/
static int
virSecuritySELinuxTransactionStart(virSecurityManagerPtr mgr)
{
bool privileged = virSecurityManagerGetPrivileged(mgr);
virSecuritySELinuxContextListPtr list;
list = virThreadLocalGet(&contextList);
if (list) {
virReportError(VIR_ERR_INTERNAL_ERROR, "%s",
_("Another relabel transaction is already started"));
return -1;
}
if (VIR_ALLOC(list) < 0)
return -1;
list->privileged = privileged;
if (virThreadLocalSet(&contextList, list) < 0) {
virReportSystemError(errno, "%s",
_("Unable to set thread local variable"));
VIR_FREE(list);
return -1;
}
return 0;
}
/**
* virSecuritySELinuxTransactionCommit:
* @mgr: security manager
* @pid: domain's PID
*
* Enters the @pid namespace (usually @pid refers to a domain) and
* performs all the sefilecon()-s on the list. Note that the
* transaction is also freed, therefore new one has to be started after
* successful return from this function. Also it is considered as error
* if there's no transaction set and this function is called.
*
* Returns: 0 on success,
* -1 otherwise.
*/
static int
virSecuritySELinuxTransactionCommit(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
pid_t pid)
{
virSecuritySELinuxContextListPtr list;
int ret = 0;
list = virThreadLocalGet(&contextList);
if (!list)
return 0;
if (virThreadLocalSet(&contextList, NULL) < 0) {
virReportSystemError(errno, "%s",
_("Unable to clear thread local variable"));
goto cleanup;
}
if (virProcessRunInMountNamespace(pid,
virSecuritySELinuxTransactionRun,
list) < 0)
goto cleanup;
ret = 0;
cleanup:
virSecuritySELinuxContextListFree(list);
return ret;
}
/**
* virSecuritySELinuxTransactionAbort:
* @mgr: security manager
*
* Cancels and frees any out standing transaction.
*/
static void
virSecuritySELinuxTransactionAbort(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED)
{
virSecuritySELinuxContextListPtr list;
list = virThreadLocalGet(&contextList);
if (!list)
return;
if (virThreadLocalSet(&contextList, NULL) < 0)
VIR_DEBUG("Unable to clear thread local variable");
virSecuritySELinuxContextListFree(list);
}
static int
virSecuritySELinuxGetProcessLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def ATTRIBUTE_UNUSED,
pid_t pid,
virSecurityLabelPtr sec)
{
security_context_t ctx;
if (getpidcon_raw(pid, &ctx) == -1) {
virReportSystemError(errno,
_("unable to get PID %d security context"),
pid);
return -1;
}
if (strlen((char *) ctx) >= VIR_SECURITY_LABEL_BUFLEN) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label exceeds "
"maximum length: %d"),
VIR_SECURITY_LABEL_BUFLEN - 1);
freecon(ctx);
return -1;
}
strcpy(sec->label, (char *) ctx);
freecon(ctx);
VIR_DEBUG("label=%s", sec->label);
sec->enforcing = security_getenforce();
if (sec->enforcing == -1) {
virReportSystemError(errno, "%s",
_("error calling security_getenforce()"));
return -1;
}
return 0;
}
/* Attempt to change the label of PATH to TCON. If OPTIONAL is true,
* return 1 if labelling was not possible. Otherwise, require a label
* change, and return 0 for success, -1 for failure. */
static int
virSecuritySELinuxSetFileconHelper(const char *path, const char *tcon,
bool optional, bool privileged)
{
security_context_t econ;
int rc;
/* Be aware that this function might run in a separate process.
* Therefore, any driver state changes would be thrown away. */
if ((rc = virSecuritySELinuxTransactionAppend(path, tcon, optional)) < 0)
return -1;
else if (rc > 0)
return 0;
VIR_INFO("Setting SELinux context on '%s' to '%s'", path, tcon);
if (setfilecon_raw(path, (VIR_SELINUX_CTX_CONST char *) tcon) < 0) {
int setfilecon_errno = errno;
if (getfilecon_raw(path, &econ) >= 0) {
if (STREQ(tcon, econ)) {
freecon(econ);
/* It's alright, there's nothing to change anyway. */
return optional ? 1 : 0;
}
freecon(econ);
}
/* If the error complaint is related to an image hosted on a (possibly
* read-only) NFS mount, or a usbfs/sysfs filesystem not supporting
* labelling, then just ignore it & hope for the best. The user
* hopefully sets one of the necessary SELinux virt_use_{nfs,usb,pci}
* boolean tunables to allow it ...
*/
VIR_WARNINGS_NO_WLOGICALOP_EQUAL_EXPR
if (setfilecon_errno != EOPNOTSUPP && setfilecon_errno != ENOTSUP &&
setfilecon_errno != EROFS) {
VIR_WARNINGS_RESET
virReportSystemError(setfilecon_errno,
_("unable to set security context '%s' on '%s'"),
tcon, path);
/* However, don't claim error if SELinux is in Enforcing mode and
* we are running as unprivileged user and we really did see EPERM.
* Otherwise we want to return error if SELinux is Enforcing. */
if (security_getenforce() == 1 && (setfilecon_errno != EPERM || privileged))
return -1;
} else {
const char *msg;
if (virFileIsSharedFSType(path, VIR_FILE_SHFS_NFS) == 1 &&
security_get_boolean_active("virt_use_nfs") != 1) {
msg = _("Setting security context '%s' on '%s' not supported. "
"Consider setting virt_use_nfs");
selinux: distinguish failure to label from request to avoid label https://bugzilla.redhat.com/show_bug.cgi?id=924153 Commit 904e05a2 (v0.9.9) added a per-<disk> seclabel element with an attribute relabel='no' in order to try and minimize the impact of shutdown delays when an NFS server disappears. The idea was that if a disk is on NFS and can't be labeled in the first place, there is no need to attempt the (no-op) relabel on domain shutdown. Unfortunately, the way this was implemented was by modifying the domain XML so that the optimization would survive libvirtd restart, but in a way that is indistinguishable from an explicit user setting. Furthermore, once the setting is turned on, libvirt avoids attempts at labeling, even for operations like snapshot or blockcopy where the chain is being extended or pivoted onto non-NFS, where SELinux labeling is once again possible. As a result, it was impossible to do a blockcopy to pivot from an NFS image file onto a local file. The solution is to separate the semantics of a chain that must not be labeled (which the user can set even on persistent domains) vs. the optimization of not attempting a relabel on cleanup (a live-only annotation), and using only the user's explicit notation rather than the optimization as the decision on whether to skip a label attempt in the first place. When upgrading an older libvirtd to a newer, an NFS volume will still attempt the relabel; but as the avoidance of a relabel was only an optimization, this shouldn't cause any problems. In the ideal future, libvirt will eventually have XML describing EVERY file in the backing chain, with each file having a separate <seclabel> element. At that point, libvirt will be able to track more closely which files need a relabel attempt at shutdown. But until we reach that point, the single <seclabel> for the entire <disk> chain is treated as a hint - when a chain has only one file, then we know it is accurate; but if the chain has more than one file, we have to attempt relabel in spite of the attribute, in case part of the chain is local and SELinux mattered for that portion of the chain. * src/conf/domain_conf.h (_virSecurityDeviceLabelDef): Add new member. * src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML): Parse it, for live images only. (virSecurityDeviceLabelDefFormat): Output it. (virDomainDiskDefParseXML, virDomainChrSourceDefParseXML) (virDomainDiskSourceDefFormat, virDomainChrDefFormat) (virDomainDiskDefFormat): Pass flags on through. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt): Honor labelskip when possible. (virSecuritySELinuxSetSecurityFileLabel): Set labelskip, not norelabel, if labeling fails. (virSecuritySELinuxSetFileconHelper): Fix indentation. * docs/formatdomain.html.in (seclabel): Document new xml. * docs/schemas/domaincommon.rng (devSeclabel): Allow it in RNG. * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.xml: * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.args: * tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-*-labelskip.xml: New test files. * tests/qemuxml2argvtest.c (mymain): Run the new tests. * tests/qemuxml2xmltest.c (mymain): Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2013-08-12 15:15:42 +00:00
if (security_getenforce() == 1)
VIR_WARN(msg, tcon, path);
else
VIR_INFO(msg, tcon, path);
} else {
VIR_INFO("Setting security context '%s' on '%s' not supported",
tcon, path);
}
if (optional)
return 1;
}
}
return 0;
}
static int
virSecuritySELinuxSetFileconOptional(virSecurityManagerPtr mgr,
const char *path, const char *tcon)
{
bool privileged = virSecurityManagerGetPrivileged(mgr);
return virSecuritySELinuxSetFileconHelper(path, tcon, true, privileged);
}
static int
virSecuritySELinuxSetFilecon(virSecurityManagerPtr mgr,
const char *path, const char *tcon)
{
bool privileged = virSecurityManagerGetPrivileged(mgr);
return virSecuritySELinuxSetFileconHelper(path, tcon, false, privileged);
}
static int
virSecuritySELinuxFSetFilecon(int fd, char *tcon)
{
security_context_t econ;
VIR_INFO("Setting SELinux context on fd %d to '%s'", fd, tcon);
if (fsetfilecon_raw(fd, tcon) < 0) {
int fsetfilecon_errno = errno;
if (fgetfilecon_raw(fd, &econ) >= 0) {
if (STREQ(tcon, econ)) {
freecon(econ);
/* It's alright, there's nothing to change anyway. */
return 0;
}
freecon(econ);
}
/* if the error complaint is related to an image hosted on
* an nfs mount, or a usbfs/sysfs filesystem not supporting
* labelling, then just ignore it & hope for the best.
* The user hopefully set one of the necessary SELinux
* virt_use_{nfs,usb,pci} boolean tunables to allow it...
*/
if (fsetfilecon_errno != EOPNOTSUPP) {
virReportSystemError(fsetfilecon_errno,
_("unable to set security context '%s' on fd %d"),
tcon, fd);
if (security_getenforce() == 1)
return -1;
} else {
VIR_INFO("Setting security context '%s' on fd %d not supported",
tcon, fd);
}
}
return 0;
}
/* Set fcon to the appropriate label for path and mode, or return -1. */
static int
getContext(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
const char *newpath, mode_t mode, security_context_t *fcon)
{
#if HAVE_SELINUX_LABEL_H
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
return selabel_lookup_raw(data->label_handle, fcon, newpath, mode);
#else
return matchpathcon(newpath, mode, fcon);
#endif
}
/* This method shouldn't raise errors, since they'll overwrite
* errors that the caller(s) are already dealing with */
static int
virSecuritySELinuxRestoreFileLabel(virSecurityManagerPtr mgr,
const char *path)
{
struct stat buf;
security_context_t fcon = NULL;
int rc = -1;
char *newpath = NULL;
char ebuf[1024];
/* Some paths are auto-generated, so let's be safe here and do
* nothing if nothing is needed.
*/
if (!path)
return 0;
VIR_INFO("Restoring SELinux context on '%s'", path);
if (virFileResolveLink(path, &newpath) < 0) {
VIR_WARN("cannot resolve symlink %s: %s", path,
virStrerror(errno, ebuf, sizeof(ebuf)));
2009-04-01 10:26:22 +00:00
goto err;
}
if (stat(newpath, &buf) != 0) {
VIR_WARN("cannot stat %s: %s", newpath,
virStrerror(errno, ebuf, sizeof(ebuf)));
2009-04-01 10:26:22 +00:00
goto err;
}
2009-04-01 10:26:22 +00:00
if (getContext(mgr, newpath, buf.st_mode, &fcon) < 0) {
/* Any user created path likely does not have a default label,
* which makes this an expected non error
*/
VIR_WARN("cannot lookup default selinux label for %s", newpath);
rc = 0;
} else {
rc = virSecuritySELinuxSetFilecon(mgr, newpath, fcon);
}
err:
freecon(fcon);
VIR_FREE(newpath);
return rc;
}
static int
virSecuritySELinuxSetInputLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainInputDefPtr input)
{
virSecurityLabelDefPtr seclabel;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (seclabel == NULL)
return 0;
switch ((virDomainInputType) input->type) {
case VIR_DOMAIN_INPUT_TYPE_PASSTHROUGH:
if (virSecuritySELinuxSetFilecon(mgr, input->source.evdev,
seclabel->imagelabel) < 0)
return -1;
break;
case VIR_DOMAIN_INPUT_TYPE_MOUSE:
case VIR_DOMAIN_INPUT_TYPE_TABLET:
case VIR_DOMAIN_INPUT_TYPE_KBD:
case VIR_DOMAIN_INPUT_TYPE_LAST:
break;
}
return 0;
}
static int
virSecuritySELinuxRestoreInputLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainInputDefPtr input)
{
int rc = 0;
virSecurityLabelDefPtr seclabel;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (seclabel == NULL)
return 0;
switch ((virDomainInputType) input->type) {
case VIR_DOMAIN_INPUT_TYPE_PASSTHROUGH:
rc = virSecuritySELinuxRestoreFileLabel(mgr, input->source.evdev);
break;
case VIR_DOMAIN_INPUT_TYPE_MOUSE:
case VIR_DOMAIN_INPUT_TYPE_TABLET:
case VIR_DOMAIN_INPUT_TYPE_KBD:
case VIR_DOMAIN_INPUT_TYPE_LAST:
break;
}
return rc;
}
static int
virSecuritySELinuxSetMemoryLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainMemoryDefPtr mem)
{
virSecurityLabelDefPtr seclabel;
switch ((virDomainMemoryModel) mem->model) {
case VIR_DOMAIN_MEMORY_MODEL_NVDIMM:
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel || !seclabel->relabel)
return 0;
if (virSecuritySELinuxSetFilecon(mgr, mem->nvdimmPath,
seclabel->imagelabel) < 0)
return -1;
break;
case VIR_DOMAIN_MEMORY_MODEL_NONE:
case VIR_DOMAIN_MEMORY_MODEL_DIMM:
case VIR_DOMAIN_MEMORY_MODEL_LAST:
break;
}
return 0;
}
static int
virSecuritySELinuxRestoreMemoryLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainMemoryDefPtr mem)
{
int ret = -1;
virSecurityLabelDefPtr seclabel;
switch ((virDomainMemoryModel) mem->model) {
case VIR_DOMAIN_MEMORY_MODEL_NVDIMM:
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel || !seclabel->relabel)
return 0;
ret = virSecuritySELinuxRestoreFileLabel(mgr, mem->nvdimmPath);
break;
case VIR_DOMAIN_MEMORY_MODEL_DIMM:
case VIR_DOMAIN_MEMORY_MODEL_NONE:
case VIR_DOMAIN_MEMORY_MODEL_LAST:
ret = 0;
break;
}
return ret;
}
static int
virSecuritySELinuxSetTPMFileLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainTPMDefPtr tpm)
{
int rc;
virSecurityLabelDefPtr seclabel;
char *cancel_path;
const char *tpmdev;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (seclabel == NULL)
return 0;
switch (tpm->type) {
case VIR_DOMAIN_TPM_TYPE_PASSTHROUGH:
tpmdev = tpm->data.passthrough.source.data.file.path;
rc = virSecuritySELinuxSetFilecon(mgr, tpmdev, seclabel->imagelabel);
if (rc < 0)
return -1;
if ((cancel_path = virTPMCreateCancelPath(tpmdev)) != NULL) {
rc = virSecuritySELinuxSetFilecon(mgr,
cancel_path,
seclabel->imagelabel);
VIR_FREE(cancel_path);
if (rc < 0) {
virSecuritySELinuxRestoreTPMFileLabelInt(mgr, def, tpm);
return -1;
}
} else {
return -1;
}
break;
case VIR_DOMAIN_TPM_TYPE_LAST:
break;
}
return 0;
}
static int
virSecuritySELinuxRestoreTPMFileLabelInt(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainTPMDefPtr tpm)
{
int rc = 0;
virSecurityLabelDefPtr seclabel;
char *cancel_path;
const char *tpmdev;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (seclabel == NULL)
return 0;
switch (tpm->type) {
case VIR_DOMAIN_TPM_TYPE_PASSTHROUGH:
tpmdev = tpm->data.passthrough.source.data.file.path;
rc = virSecuritySELinuxRestoreFileLabel(mgr, tpmdev);
if ((cancel_path = virTPMCreateCancelPath(tpmdev)) != NULL) {
if (virSecuritySELinuxRestoreFileLabel(mgr, cancel_path) < 0)
rc = -1;
VIR_FREE(cancel_path);
}
break;
case VIR_DOMAIN_TPM_TYPE_LAST:
break;
}
return rc;
}
static int
virSecuritySELinuxRestoreImageLabelInt(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virStorageSourcePtr src,
bool migrated)
{
virSecurityLabelDefPtr seclabel;
virSecurityDeviceLabelDefPtr disk_seclabel;
if (!src->path || !virStorageSourceIsLocalStorage(src))
return 0;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (seclabel == NULL)
return 0;
disk_seclabel = virStorageSourceGetSecurityLabelDef(src,
SECURITY_SELINUX_NAME);
if (!seclabel->relabel || (disk_seclabel && !disk_seclabel->relabel))
return 0;
selinux: distinguish failure to label from request to avoid label https://bugzilla.redhat.com/show_bug.cgi?id=924153 Commit 904e05a2 (v0.9.9) added a per-<disk> seclabel element with an attribute relabel='no' in order to try and minimize the impact of shutdown delays when an NFS server disappears. The idea was that if a disk is on NFS and can't be labeled in the first place, there is no need to attempt the (no-op) relabel on domain shutdown. Unfortunately, the way this was implemented was by modifying the domain XML so that the optimization would survive libvirtd restart, but in a way that is indistinguishable from an explicit user setting. Furthermore, once the setting is turned on, libvirt avoids attempts at labeling, even for operations like snapshot or blockcopy where the chain is being extended or pivoted onto non-NFS, where SELinux labeling is once again possible. As a result, it was impossible to do a blockcopy to pivot from an NFS image file onto a local file. The solution is to separate the semantics of a chain that must not be labeled (which the user can set even on persistent domains) vs. the optimization of not attempting a relabel on cleanup (a live-only annotation), and using only the user's explicit notation rather than the optimization as the decision on whether to skip a label attempt in the first place. When upgrading an older libvirtd to a newer, an NFS volume will still attempt the relabel; but as the avoidance of a relabel was only an optimization, this shouldn't cause any problems. In the ideal future, libvirt will eventually have XML describing EVERY file in the backing chain, with each file having a separate <seclabel> element. At that point, libvirt will be able to track more closely which files need a relabel attempt at shutdown. But until we reach that point, the single <seclabel> for the entire <disk> chain is treated as a hint - when a chain has only one file, then we know it is accurate; but if the chain has more than one file, we have to attempt relabel in spite of the attribute, in case part of the chain is local and SELinux mattered for that portion of the chain. * src/conf/domain_conf.h (_virSecurityDeviceLabelDef): Add new member. * src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML): Parse it, for live images only. (virSecurityDeviceLabelDefFormat): Output it. (virDomainDiskDefParseXML, virDomainChrSourceDefParseXML) (virDomainDiskSourceDefFormat, virDomainChrDefFormat) (virDomainDiskDefFormat): Pass flags on through. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt): Honor labelskip when possible. (virSecuritySELinuxSetSecurityFileLabel): Set labelskip, not norelabel, if labeling fails. (virSecuritySELinuxSetFileconHelper): Fix indentation. * docs/formatdomain.html.in (seclabel): Document new xml. * docs/schemas/domaincommon.rng (devSeclabel): Allow it in RNG. * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.xml: * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.args: * tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-*-labelskip.xml: New test files. * tests/qemuxml2argvtest.c (mymain): Run the new tests. * tests/qemuxml2xmltest.c (mymain): Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2013-08-12 15:15:42 +00:00
/* If labelskip is true and there are no backing files, then we
* know it is safe to skip the restore. FIXME - backing files should
* be tracked in domain XML, at which point labelskip should be a
* per-file attribute instead of a disk attribute. */
selinux: distinguish failure to label from request to avoid label https://bugzilla.redhat.com/show_bug.cgi?id=924153 Commit 904e05a2 (v0.9.9) added a per-<disk> seclabel element with an attribute relabel='no' in order to try and minimize the impact of shutdown delays when an NFS server disappears. The idea was that if a disk is on NFS and can't be labeled in the first place, there is no need to attempt the (no-op) relabel on domain shutdown. Unfortunately, the way this was implemented was by modifying the domain XML so that the optimization would survive libvirtd restart, but in a way that is indistinguishable from an explicit user setting. Furthermore, once the setting is turned on, libvirt avoids attempts at labeling, even for operations like snapshot or blockcopy where the chain is being extended or pivoted onto non-NFS, where SELinux labeling is once again possible. As a result, it was impossible to do a blockcopy to pivot from an NFS image file onto a local file. The solution is to separate the semantics of a chain that must not be labeled (which the user can set even on persistent domains) vs. the optimization of not attempting a relabel on cleanup (a live-only annotation), and using only the user's explicit notation rather than the optimization as the decision on whether to skip a label attempt in the first place. When upgrading an older libvirtd to a newer, an NFS volume will still attempt the relabel; but as the avoidance of a relabel was only an optimization, this shouldn't cause any problems. In the ideal future, libvirt will eventually have XML describing EVERY file in the backing chain, with each file having a separate <seclabel> element. At that point, libvirt will be able to track more closely which files need a relabel attempt at shutdown. But until we reach that point, the single <seclabel> for the entire <disk> chain is treated as a hint - when a chain has only one file, then we know it is accurate; but if the chain has more than one file, we have to attempt relabel in spite of the attribute, in case part of the chain is local and SELinux mattered for that portion of the chain. * src/conf/domain_conf.h (_virSecurityDeviceLabelDef): Add new member. * src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML): Parse it, for live images only. (virSecurityDeviceLabelDefFormat): Output it. (virDomainDiskDefParseXML, virDomainChrSourceDefParseXML) (virDomainDiskSourceDefFormat, virDomainChrDefFormat) (virDomainDiskDefFormat): Pass flags on through. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt): Honor labelskip when possible. (virSecuritySELinuxSetSecurityFileLabel): Set labelskip, not norelabel, if labeling fails. (virSecuritySELinuxSetFileconHelper): Fix indentation. * docs/formatdomain.html.in (seclabel): Document new xml. * docs/schemas/domaincommon.rng (devSeclabel): Allow it in RNG. * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.xml: * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.args: * tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-*-labelskip.xml: New test files. * tests/qemuxml2argvtest.c (mymain): Run the new tests. * tests/qemuxml2xmltest.c (mymain): Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2013-08-12 15:15:42 +00:00
if (disk_seclabel && disk_seclabel->labelskip &&
!src->backingStore)
selinux: distinguish failure to label from request to avoid label https://bugzilla.redhat.com/show_bug.cgi?id=924153 Commit 904e05a2 (v0.9.9) added a per-<disk> seclabel element with an attribute relabel='no' in order to try and minimize the impact of shutdown delays when an NFS server disappears. The idea was that if a disk is on NFS and can't be labeled in the first place, there is no need to attempt the (no-op) relabel on domain shutdown. Unfortunately, the way this was implemented was by modifying the domain XML so that the optimization would survive libvirtd restart, but in a way that is indistinguishable from an explicit user setting. Furthermore, once the setting is turned on, libvirt avoids attempts at labeling, even for operations like snapshot or blockcopy where the chain is being extended or pivoted onto non-NFS, where SELinux labeling is once again possible. As a result, it was impossible to do a blockcopy to pivot from an NFS image file onto a local file. The solution is to separate the semantics of a chain that must not be labeled (which the user can set even on persistent domains) vs. the optimization of not attempting a relabel on cleanup (a live-only annotation), and using only the user's explicit notation rather than the optimization as the decision on whether to skip a label attempt in the first place. When upgrading an older libvirtd to a newer, an NFS volume will still attempt the relabel; but as the avoidance of a relabel was only an optimization, this shouldn't cause any problems. In the ideal future, libvirt will eventually have XML describing EVERY file in the backing chain, with each file having a separate <seclabel> element. At that point, libvirt will be able to track more closely which files need a relabel attempt at shutdown. But until we reach that point, the single <seclabel> for the entire <disk> chain is treated as a hint - when a chain has only one file, then we know it is accurate; but if the chain has more than one file, we have to attempt relabel in spite of the attribute, in case part of the chain is local and SELinux mattered for that portion of the chain. * src/conf/domain_conf.h (_virSecurityDeviceLabelDef): Add new member. * src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML): Parse it, for live images only. (virSecurityDeviceLabelDefFormat): Output it. (virDomainDiskDefParseXML, virDomainChrSourceDefParseXML) (virDomainDiskSourceDefFormat, virDomainChrDefFormat) (virDomainDiskDefFormat): Pass flags on through. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt): Honor labelskip when possible. (virSecuritySELinuxSetSecurityFileLabel): Set labelskip, not norelabel, if labeling fails. (virSecuritySELinuxSetFileconHelper): Fix indentation. * docs/formatdomain.html.in (seclabel): Document new xml. * docs/schemas/domaincommon.rng (devSeclabel): Allow it in RNG. * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.xml: * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.args: * tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-*-labelskip.xml: New test files. * tests/qemuxml2argvtest.c (mymain): Run the new tests. * tests/qemuxml2xmltest.c (mymain): Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2013-08-12 15:15:42 +00:00
return 0;
/* Don't restore labels on readonly/shared disks, because other VMs may
* still be accessing these. Alternatively we could iterate over all
* running domains and try to figure out if it is in use, but this would
* not work for clustered filesystems, since we can't see running VMs using
* the file on other nodes. Safest bet is thus to skip the restore step. */
if (src->readonly || src->shared)
return 0;
/* If we have a shared FS and are doing migration, we must not change
* ownership, because that kills access on the destination host which is
* sub-optimal for the guest VM's I/O attempts :-) */
if (migrated) {
int rc = virFileIsSharedFS(src->path);
if (rc < 0)
return -1;
if (rc == 1) {
VIR_DEBUG("Skipping image label restore on %s because FS is shared",
src->path);
return 0;
}
}
return virSecuritySELinuxRestoreFileLabel(mgr, src->path);
}
static int
virSecuritySELinuxRestoreDiskLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainDiskDefPtr disk)
{
return virSecuritySELinuxRestoreImageLabelInt(mgr, def, disk->src,
false);
}
static int
virSecuritySELinuxRestoreImageLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virStorageSourcePtr src)
{
return virSecuritySELinuxRestoreImageLabelInt(mgr, def, src, false);
}
static int
virSecuritySELinuxSetImageLabelInternal(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virStorageSourcePtr src,
bool first)
{
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
virSecurityLabelDefPtr secdef;
virSecurityDeviceLabelDefPtr disk_seclabel;
int ret;
if (!src->path || !virStorageSourceIsLocalStorage(src))
return 0;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel)
return 0;
disk_seclabel = virStorageSourceGetSecurityLabelDef(src,
SECURITY_SELINUX_NAME);
if (disk_seclabel && !disk_seclabel->relabel)
return 0;
if (disk_seclabel && disk_seclabel->relabel && disk_seclabel->label) {
ret = virSecuritySELinuxSetFilecon(mgr, src->path, disk_seclabel->label);
} else if (first) {
if (src->shared) {
ret = virSecuritySELinuxSetFileconOptional(mgr,
src->path,
data->file_context);
} else if (src->readonly) {
ret = virSecuritySELinuxSetFileconOptional(mgr,
src->path,
data->content_context);
} else if (secdef->imagelabel) {
ret = virSecuritySELinuxSetFileconOptional(mgr,
src->path,
secdef->imagelabel);
} else {
ret = 0;
}
} else {
ret = virSecuritySELinuxSetFileconOptional(mgr,
src->path,
data->content_context);
}
if (ret == 1 && !disk_seclabel) {
/* If we failed to set a label, but virt_use_nfs let us
* proceed anyway, then we don't need to relabel later. */
disk_seclabel = virSecurityDeviceLabelDefNew(SECURITY_SELINUX_NAME);
if (!disk_seclabel)
return -1;
selinux: distinguish failure to label from request to avoid label https://bugzilla.redhat.com/show_bug.cgi?id=924153 Commit 904e05a2 (v0.9.9) added a per-<disk> seclabel element with an attribute relabel='no' in order to try and minimize the impact of shutdown delays when an NFS server disappears. The idea was that if a disk is on NFS and can't be labeled in the first place, there is no need to attempt the (no-op) relabel on domain shutdown. Unfortunately, the way this was implemented was by modifying the domain XML so that the optimization would survive libvirtd restart, but in a way that is indistinguishable from an explicit user setting. Furthermore, once the setting is turned on, libvirt avoids attempts at labeling, even for operations like snapshot or blockcopy where the chain is being extended or pivoted onto non-NFS, where SELinux labeling is once again possible. As a result, it was impossible to do a blockcopy to pivot from an NFS image file onto a local file. The solution is to separate the semantics of a chain that must not be labeled (which the user can set even on persistent domains) vs. the optimization of not attempting a relabel on cleanup (a live-only annotation), and using only the user's explicit notation rather than the optimization as the decision on whether to skip a label attempt in the first place. When upgrading an older libvirtd to a newer, an NFS volume will still attempt the relabel; but as the avoidance of a relabel was only an optimization, this shouldn't cause any problems. In the ideal future, libvirt will eventually have XML describing EVERY file in the backing chain, with each file having a separate <seclabel> element. At that point, libvirt will be able to track more closely which files need a relabel attempt at shutdown. But until we reach that point, the single <seclabel> for the entire <disk> chain is treated as a hint - when a chain has only one file, then we know it is accurate; but if the chain has more than one file, we have to attempt relabel in spite of the attribute, in case part of the chain is local and SELinux mattered for that portion of the chain. * src/conf/domain_conf.h (_virSecurityDeviceLabelDef): Add new member. * src/conf/domain_conf.c (virSecurityDeviceLabelDefParseXML): Parse it, for live images only. (virSecurityDeviceLabelDefFormat): Output it. (virDomainDiskDefParseXML, virDomainChrSourceDefParseXML) (virDomainDiskSourceDefFormat, virDomainChrDefFormat) (virDomainDiskDefFormat): Pass flags on through. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt): Honor labelskip when possible. (virSecuritySELinuxSetSecurityFileLabel): Set labelskip, not norelabel, if labeling fails. (virSecuritySELinuxSetFileconHelper): Fix indentation. * docs/formatdomain.html.in (seclabel): Document new xml. * docs/schemas/domaincommon.rng (devSeclabel): Allow it in RNG. * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.xml: * tests/qemuxml2argvdata/qemuxml2argv-seclabel-*-labelskip.args: * tests/qemuxml2xmloutdata/qemuxml2xmlout-seclabel-*-labelskip.xml: New test files. * tests/qemuxml2argvtest.c (mymain): Run the new tests. * tests/qemuxml2xmltest.c (mymain): Likewise. Signed-off-by: Eric Blake <eblake@redhat.com>
2013-08-12 15:15:42 +00:00
disk_seclabel->labelskip = true;
if (VIR_APPEND_ELEMENT(src->seclabels, src->nseclabels,
disk_seclabel) < 0) {
virSecurityDeviceLabelDefFree(disk_seclabel);
return -1;
}
ret = 0;
}
return ret;
}
static int
virSecuritySELinuxSetImageLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virStorageSourcePtr src)
{
return virSecuritySELinuxSetImageLabelInternal(mgr, def, src, true);
}
static int
virSecuritySELinuxSetDiskLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainDiskDefPtr disk)
{
bool first = true;
virStorageSourcePtr next;
for (next = disk->src; next; next = next->backingStore) {
if (virSecuritySELinuxSetImageLabelInternal(mgr, def, next, first) < 0)
return -1;
first = false;
}
return 0;
}
static int
virSecuritySELinuxSetHostdevLabelHelper(const char *file, void *opaque)
{
virSecurityLabelDefPtr secdef;
virSecuritySELinuxCallbackDataPtr data = opaque;
virSecurityManagerPtr mgr = data->mgr;
virDomainDefPtr def = data->def;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
return 0;
return virSecuritySELinuxSetFilecon(mgr, file, secdef->imagelabel);
}
static int
virSecuritySELinuxSetPCILabel(virPCIDevicePtr dev ATTRIBUTE_UNUSED,
const char *file, void *opaque)
{
return virSecuritySELinuxSetHostdevLabelHelper(file, opaque);
}
static int
virSecuritySELinuxSetUSBLabel(virUSBDevicePtr dev ATTRIBUTE_UNUSED,
const char *file, void *opaque)
{
return virSecuritySELinuxSetHostdevLabelHelper(file, opaque);
}
static int
virSecuritySELinuxSetSCSILabel(virSCSIDevicePtr dev,
const char *file, void *opaque)
{
virSecurityLabelDefPtr secdef;
virSecuritySELinuxCallbackDataPtr ptr = opaque;
virSecurityManagerPtr mgr = ptr->mgr;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
secdef = virDomainDefGetSecurityLabelDef(ptr->def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
return 0;
if (virSCSIDeviceGetShareable(dev))
return virSecuritySELinuxSetFileconOptional(mgr, file,
data->file_context);
else if (virSCSIDeviceGetReadonly(dev))
return virSecuritySELinuxSetFileconOptional(mgr, file,
data->content_context);
else
return virSecuritySELinuxSetFileconOptional(mgr, file,
secdef->imagelabel);
}
static int
virSecuritySELinuxSetHostLabel(virSCSIVHostDevicePtr dev ATTRIBUTE_UNUSED,
const char *file, void *opaque)
{
return virSecuritySELinuxSetHostdevLabelHelper(file, opaque);
}
static int
virSecuritySELinuxSetHostdevSubsysLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainHostdevDefPtr dev,
const char *vroot)
{
virDomainHostdevSubsysUSBPtr usbsrc = &dev->source.subsys.u.usb;
virDomainHostdevSubsysPCIPtr pcisrc = &dev->source.subsys.u.pci;
virDomainHostdevSubsysSCSIPtr scsisrc = &dev->source.subsys.u.scsi;
virDomainHostdevSubsysSCSIVHostPtr hostsrc = &dev->source.subsys.u.scsi_host;
virDomainHostdevSubsysMediatedDevPtr mdevsrc = &dev->source.subsys.u.mdev;
virSecuritySELinuxCallbackData data = {.mgr = mgr, .def = def};
int ret = -1;
/* Like virSecuritySELinuxSetImageLabelInternal() for a networked
* disk, do nothing for an iSCSI hostdev
*/
if (dev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI &&
scsisrc->protocol == VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI)
return 0;
switch ((virDomainHostdevSubsysType) dev->source.subsys.type) {
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB: {
virUSBDevicePtr usb;
if (dev->missing)
return 0;
usb = virUSBDeviceNew(usbsrc->bus,
usbsrc->device,
vroot);
if (!usb)
goto done;
ret = virUSBDeviceFileIterate(usb, virSecuritySELinuxSetUSBLabel, &data);
virUSBDeviceFree(usb);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI: {
virPCIDevicePtr pci =
virPCIDeviceNew(pcisrc->addr.domain, pcisrc->addr.bus,
pcisrc->addr.slot, pcisrc->addr.function);
if (!pci)
goto done;
if (pcisrc->backend == VIR_DOMAIN_HOSTDEV_PCI_BACKEND_VFIO) {
char *vfioGroupDev = virPCIDeviceGetIOMMUGroupDev(pci);
if (!vfioGroupDev) {
virPCIDeviceFree(pci);
goto done;
}
ret = virSecuritySELinuxSetPCILabel(pci, vfioGroupDev, &data);
VIR_FREE(vfioGroupDev);
} else {
ret = virPCIDeviceFileIterate(pci, virSecuritySELinuxSetPCILabel, &data);
}
virPCIDeviceFree(pci);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI: {
virDomainHostdevSubsysSCSIHostPtr scsihostsrc = &scsisrc->u.host;
virSCSIDevicePtr scsi =
virSCSIDeviceNew(NULL,
scsihostsrc->adapter, scsihostsrc->bus,
scsihostsrc->target, scsihostsrc->unit,
dev->readonly, dev->shareable);
if (!scsi)
goto done;
ret = virSCSIDeviceFileIterate(scsi,
virSecuritySELinuxSetSCSILabel,
&data);
virSCSIDeviceFree(scsi);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST: {
virSCSIVHostDevicePtr host = virSCSIVHostDeviceNew(hostsrc->wwpn);
if (!host)
goto done;
ret = virSCSIVHostDeviceFileIterate(host,
virSecuritySELinuxSetHostLabel,
&data);
virSCSIVHostDeviceFree(host);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV: {
char *vfiodev = NULL;
if (!(vfiodev = virMediatedDeviceGetIOMMUGroupDev(mdevsrc->uuidstr)))
goto done;
ret = virSecuritySELinuxSetHostdevLabelHelper(vfiodev, &data);
VIR_FREE(vfiodev);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_LAST:
ret = 0;
break;
}
done:
return ret;
}
static int
virSecuritySELinuxSetHostdevCapsLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainHostdevDefPtr dev,
const char *vroot)
{
int ret = -1;
virSecurityLabelDefPtr secdef;
char *path;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
return 0;
switch (dev->source.caps.type) {
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_STORAGE: {
if (vroot) {
if (virAsprintf(&path, "%s/%s", vroot,
dev->source.caps.u.storage.block) < 0)
return -1;
} else {
if (VIR_STRDUP(path, dev->source.caps.u.storage.block) < 0)
return -1;
}
ret = virSecuritySELinuxSetFilecon(mgr, path, secdef->imagelabel);
VIR_FREE(path);
break;
}
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_MISC: {
if (vroot) {
if (virAsprintf(&path, "%s/%s", vroot,
dev->source.caps.u.misc.chardev) < 0)
return -1;
} else {
if (VIR_STRDUP(path, dev->source.caps.u.misc.chardev) < 0)
return -1;
}
ret = virSecuritySELinuxSetFilecon(mgr, path, secdef->imagelabel);
VIR_FREE(path);
break;
}
default:
ret = 0;
break;
}
return ret;
}
static int
virSecuritySELinuxSetHostdevLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainHostdevDefPtr dev,
const char *vroot)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel)
return 0;
switch (dev->mode) {
case VIR_DOMAIN_HOSTDEV_MODE_SUBSYS:
return virSecuritySELinuxSetHostdevSubsysLabel(mgr, def, dev, vroot);
case VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES:
return virSecuritySELinuxSetHostdevCapsLabel(mgr, def, dev, vroot);
default:
return 0;
}
}
static int
virSecuritySELinuxRestorePCILabel(virPCIDevicePtr dev ATTRIBUTE_UNUSED,
const char *file,
void *opaque)
{
virSecurityManagerPtr mgr = opaque;
return virSecuritySELinuxRestoreFileLabel(mgr, file);
}
static int
virSecuritySELinuxRestoreUSBLabel(virUSBDevicePtr dev ATTRIBUTE_UNUSED,
const char *file,
void *opaque)
{
virSecurityManagerPtr mgr = opaque;
return virSecuritySELinuxRestoreFileLabel(mgr, file);
}
static int
virSecuritySELinuxRestoreSCSILabel(virSCSIDevicePtr dev,
const char *file,
void *opaque)
{
virSecurityManagerPtr mgr = opaque;
/* Don't restore labels on a shareable or readonly hostdev, because
* other VMs may still be accessing.
*/
if (virSCSIDeviceGetShareable(dev) || virSCSIDeviceGetReadonly(dev))
return 0;
return virSecuritySELinuxRestoreFileLabel(mgr, file);
}
static int
virSecuritySELinuxRestoreHostLabel(virSCSIVHostDevicePtr dev ATTRIBUTE_UNUSED,
const char *file,
void *opaque)
{
virSecurityManagerPtr mgr = opaque;
return virSecuritySELinuxRestoreFileLabel(mgr, file);
}
static int
virSecuritySELinuxRestoreHostdevSubsysLabel(virSecurityManagerPtr mgr,
virDomainHostdevDefPtr dev,
const char *vroot)
{
virDomainHostdevSubsysUSBPtr usbsrc = &dev->source.subsys.u.usb;
virDomainHostdevSubsysPCIPtr pcisrc = &dev->source.subsys.u.pci;
virDomainHostdevSubsysSCSIPtr scsisrc = &dev->source.subsys.u.scsi;
virDomainHostdevSubsysSCSIVHostPtr hostsrc = &dev->source.subsys.u.scsi_host;
virDomainHostdevSubsysMediatedDevPtr mdevsrc = &dev->source.subsys.u.mdev;
int ret = -1;
/* Like virSecuritySELinuxRestoreImageLabelInt() for a networked
* disk, do nothing for an iSCSI hostdev
*/
if (dev->source.subsys.type == VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI &&
scsisrc->protocol == VIR_DOMAIN_HOSTDEV_SCSI_PROTOCOL_TYPE_ISCSI)
return 0;
switch ((virDomainHostdevSubsysType) dev->source.subsys.type) {
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_USB: {
virUSBDevicePtr usb;
if (dev->missing)
return 0;
usb = virUSBDeviceNew(usbsrc->bus,
usbsrc->device,
vroot);
if (!usb)
goto done;
ret = virUSBDeviceFileIterate(usb, virSecuritySELinuxRestoreUSBLabel, mgr);
virUSBDeviceFree(usb);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_PCI: {
virPCIDevicePtr pci =
virPCIDeviceNew(pcisrc->addr.domain, pcisrc->addr.bus,
pcisrc->addr.slot, pcisrc->addr.function);
if (!pci)
goto done;
if (pcisrc->backend == VIR_DOMAIN_HOSTDEV_PCI_BACKEND_VFIO) {
char *vfioGroupDev = virPCIDeviceGetIOMMUGroupDev(pci);
if (!vfioGroupDev) {
virPCIDeviceFree(pci);
goto done;
}
ret = virSecuritySELinuxRestorePCILabel(pci, vfioGroupDev, mgr);
VIR_FREE(vfioGroupDev);
} else {
ret = virPCIDeviceFileIterate(pci, virSecuritySELinuxRestorePCILabel, mgr);
}
virPCIDeviceFree(pci);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI: {
virDomainHostdevSubsysSCSIHostPtr scsihostsrc = &scsisrc->u.host;
virSCSIDevicePtr scsi =
virSCSIDeviceNew(NULL,
scsihostsrc->adapter, scsihostsrc->bus,
scsihostsrc->target, scsihostsrc->unit,
dev->readonly, dev->shareable);
2015-11-20 09:29:21 +00:00
if (!scsi)
goto done;
ret = virSCSIDeviceFileIterate(scsi, virSecuritySELinuxRestoreSCSILabel, mgr);
2015-11-20 09:29:21 +00:00
virSCSIDeviceFree(scsi);
2015-11-20 09:29:21 +00:00
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_SCSI_HOST: {
virSCSIVHostDevicePtr host = virSCSIVHostDeviceNew(hostsrc->wwpn);
if (!host)
goto done;
ret = virSCSIVHostDeviceFileIterate(host,
virSecuritySELinuxRestoreHostLabel,
mgr);
virSCSIVHostDeviceFree(host);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_MDEV: {
char *vfiodev = NULL;
if (!(vfiodev = virMediatedDeviceGetIOMMUGroupDev(mdevsrc->uuidstr)))
goto done;
ret = virSecuritySELinuxRestoreFileLabel(mgr, vfiodev);
VIR_FREE(vfiodev);
break;
}
case VIR_DOMAIN_HOSTDEV_SUBSYS_TYPE_LAST:
ret = 0;
break;
}
done:
return ret;
}
static int
virSecuritySELinuxRestoreHostdevCapsLabel(virSecurityManagerPtr mgr,
virDomainHostdevDefPtr dev,
const char *vroot)
{
int ret = -1;
char *path;
switch (dev->source.caps.type) {
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_STORAGE: {
if (vroot) {
if (virAsprintf(&path, "%s/%s", vroot,
dev->source.caps.u.storage.block) < 0)
return -1;
} else {
if (VIR_STRDUP(path, dev->source.caps.u.storage.block) < 0)
return -1;
}
ret = virSecuritySELinuxRestoreFileLabel(mgr, path);
VIR_FREE(path);
break;
}
case VIR_DOMAIN_HOSTDEV_CAPS_TYPE_MISC: {
if (vroot) {
if (virAsprintf(&path, "%s/%s", vroot,
dev->source.caps.u.misc.chardev) < 0)
return -1;
} else {
if (VIR_STRDUP(path, dev->source.caps.u.misc.chardev) < 0)
return -1;
}
ret = virSecuritySELinuxRestoreFileLabel(mgr, path);
VIR_FREE(path);
break;
}
default:
ret = 0;
break;
}
return ret;
}
static int
virSecuritySELinuxRestoreHostdevLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainHostdevDefPtr dev,
const char *vroot)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel)
return 0;
switch (dev->mode) {
case VIR_DOMAIN_HOSTDEV_MODE_SUBSYS:
return virSecuritySELinuxRestoreHostdevSubsysLabel(mgr, dev, vroot);
case VIR_DOMAIN_HOSTDEV_MODE_CAPABILITIES:
return virSecuritySELinuxRestoreHostdevCapsLabel(mgr, dev, vroot);
default:
return 0;
}
}
static int
virSecuritySELinuxSetChardevLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainChrSourceDefPtr dev_source,
bool chardevStdioLogd)
{
virSecurityLabelDefPtr seclabel;
virSecurityDeviceLabelDefPtr chr_seclabel = NULL;
char *imagelabel = NULL;
char *in = NULL, *out = NULL;
int ret = -1;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel || !seclabel->relabel)
return 0;
chr_seclabel = virDomainChrSourceDefGetSecurityLabelDef(dev_source,
SECURITY_SELINUX_NAME);
if (chr_seclabel && !chr_seclabel->relabel)
return 0;
if (!chr_seclabel &&
dev_source->type == VIR_DOMAIN_CHR_TYPE_FILE &&
chardevStdioLogd)
return 0;
if (chr_seclabel)
imagelabel = chr_seclabel->label;
if (!imagelabel)
imagelabel = seclabel->imagelabel;
switch (dev_source->type) {
case VIR_DOMAIN_CHR_TYPE_DEV:
case VIR_DOMAIN_CHR_TYPE_FILE:
ret = virSecuritySELinuxSetFilecon(mgr,
dev_source->data.file.path,
imagelabel);
break;
case VIR_DOMAIN_CHR_TYPE_UNIX:
if (!dev_source->data.nix.listen) {
if (virSecuritySELinuxSetFilecon(mgr,
dev_source->data.nix.path,
imagelabel) < 0)
goto done;
}
ret = 0;
break;
case VIR_DOMAIN_CHR_TYPE_PIPE:
if ((virAsprintf(&in, "%s.in", dev_source->data.file.path) < 0) ||
(virAsprintf(&out, "%s.out", dev_source->data.file.path) < 0))
goto done;
if (virFileExists(in) && virFileExists(out)) {
if ((virSecuritySELinuxSetFilecon(mgr, in, imagelabel) < 0) ||
(virSecuritySELinuxSetFilecon(mgr, out, imagelabel) < 0)) {
goto done;
}
} else if (virSecuritySELinuxSetFilecon(mgr,
dev_source->data.file.path,
imagelabel) < 0) {
goto done;
}
ret = 0;
break;
default:
ret = 0;
break;
}
done:
VIR_FREE(in);
VIR_FREE(out);
return ret;
}
static int
virSecuritySELinuxRestoreChardevLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
virDomainChrSourceDefPtr dev_source,
bool chardevStdioLogd)
{
virSecurityLabelDefPtr seclabel;
virSecurityDeviceLabelDefPtr chr_seclabel = NULL;
char *in = NULL, *out = NULL;
int ret = -1;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel || !seclabel->relabel)
return 0;
chr_seclabel = virDomainChrSourceDefGetSecurityLabelDef(dev_source,
SECURITY_SELINUX_NAME);
if (chr_seclabel && !chr_seclabel->relabel)
return 0;
if (!chr_seclabel &&
dev_source->type == VIR_DOMAIN_CHR_TYPE_FILE &&
chardevStdioLogd)
return 0;
switch (dev_source->type) {
case VIR_DOMAIN_CHR_TYPE_DEV:
case VIR_DOMAIN_CHR_TYPE_FILE:
if (virSecuritySELinuxRestoreFileLabel(mgr, dev_source->data.file.path) < 0)
goto done;
ret = 0;
break;
case VIR_DOMAIN_CHR_TYPE_UNIX:
if (!dev_source->data.nix.listen) {
if (virSecuritySELinuxRestoreFileLabel(mgr, dev_source->data.file.path) < 0)
goto done;
}
ret = 0;
break;
case VIR_DOMAIN_CHR_TYPE_PIPE:
if ((virAsprintf(&out, "%s.out", dev_source->data.file.path) < 0) ||
(virAsprintf(&in, "%s.in", dev_source->data.file.path) < 0))
goto done;
if (virFileExists(in) && virFileExists(out)) {
if ((virSecuritySELinuxRestoreFileLabel(mgr, out) < 0) ||
(virSecuritySELinuxRestoreFileLabel(mgr, in) < 0)) {
goto done;
}
} else if (virSecuritySELinuxRestoreFileLabel(mgr, dev_source->data.file.path) < 0) {
goto done;
}
ret = 0;
break;
default:
ret = 0;
break;
}
done:
VIR_FREE(in);
VIR_FREE(out);
return ret;
}
struct _virSecuritySELinuxChardevCallbackData {
virSecurityManagerPtr mgr;
bool chardevStdioLogd;
};
static int
virSecuritySELinuxRestoreSecurityChardevCallback(virDomainDefPtr def,
virDomainChrDefPtr dev ATTRIBUTE_UNUSED,
void *opaque)
{
struct _virSecuritySELinuxChardevCallbackData *data = opaque;
return virSecuritySELinuxRestoreChardevLabel(data->mgr, def, dev->source,
data->chardevStdioLogd);
}
static int
virSecuritySELinuxRestoreSecuritySmartcardCallback(virDomainDefPtr def,
virDomainSmartcardDefPtr dev,
void *opaque)
{
virSecurityManagerPtr mgr = opaque;
const char *database;
switch (dev->type) {
case VIR_DOMAIN_SMARTCARD_TYPE_HOST:
break;
case VIR_DOMAIN_SMARTCARD_TYPE_HOST_CERTIFICATES:
database = dev->data.cert.database;
if (!database)
database = VIR_DOMAIN_SMARTCARD_DEFAULT_DATABASE;
return virSecuritySELinuxRestoreFileLabel(mgr, database);
case VIR_DOMAIN_SMARTCARD_TYPE_PASSTHROUGH:
return virSecuritySELinuxRestoreChardevLabel(mgr, def,
dev->data.passthru, false);
default:
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unknown smartcard type %d"),
dev->type);
return -1;
}
return 0;
}
static const char *
virSecuritySELinuxGetBaseLabel(virSecurityManagerPtr mgr, int virtType)
{
virSecuritySELinuxDataPtr priv = virSecurityManagerGetPrivateData(mgr);
if (virtType == VIR_DOMAIN_VIRT_QEMU && priv->alt_domain_context)
return priv->alt_domain_context;
else
return priv->domain_context;
}
static int
virSecuritySELinuxRestoreAllLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
bool migrated,
bool chardevStdioLogd)
{
virSecurityLabelDefPtr secdef;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
size_t i;
int rc = 0;
VIR_DEBUG("Restoring security label on %s", def->name);
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel || data->skipAllLabel)
return 0;
if (def->tpm) {
if (virSecuritySELinuxRestoreTPMFileLabelInt(mgr, def, def->tpm) < 0)
rc = -1;
}
for (i = 0; i < def->nhostdevs; i++) {
if (virSecuritySELinuxRestoreHostdevLabel(mgr,
def,
def->hostdevs[i],
NULL) < 0)
rc = -1;
}
for (i = 0; i < def->ninputs; i++) {
if (virSecuritySELinuxRestoreInputLabel(mgr, def, def->inputs[i]) < 0)
rc = -1;
}
for (i = 0; i < def->nmems; i++) {
if (virSecuritySELinuxRestoreMemoryLabel(mgr, def, def->mems[i]) < 0)
return -1;
}
for (i = 0; i < def->ndisks; i++) {
virDomainDiskDefPtr disk = def->disks[i];
if (virSecuritySELinuxRestoreImageLabelInt(mgr, def, disk->src,
migrated) < 0)
rc = -1;
}
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
struct _virSecuritySELinuxChardevCallbackData chardevData = {
.mgr = mgr,
.chardevStdioLogd = chardevStdioLogd
};
if (virDomainChrDefForeach(def,
false,
virSecuritySELinuxRestoreSecurityChardevCallback,
&chardevData) < 0)
rc = -1;
if (virDomainSmartcardDefForeach(def,
false,
virSecuritySELinuxRestoreSecuritySmartcardCallback,
mgr) < 0)
rc = -1;
if (def->os.loader && def->os.loader->nvram &&
virSecuritySELinuxRestoreFileLabel(mgr, def->os.loader->nvram) < 0)
rc = -1;
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return rc;
}
static int
virSecuritySELinuxReleaseLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def)
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
return 0;
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
if (secdef->type == VIR_DOMAIN_SECLABEL_DYNAMIC) {
if (secdef->label != NULL) {
context_t con = context_new(secdef->label);
if (con) {
virSecuritySELinuxMCSRemove(mgr, context_range_get(con));
context_free(con);
}
}
VIR_FREE(secdef->label);
if (!secdef->baselabel)
VIR_FREE(secdef->model);
}
VIR_FREE(secdef->imagelabel);
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return 0;
}
static int
virSecuritySELinuxSetSavedStateLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
const char *savefile)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel)
return 0;
return virSecuritySELinuxSetFilecon(mgr, savefile, secdef->imagelabel);
}
static int
virSecuritySELinuxRestoreSavedStateLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
const char *savefile)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->relabel)
return 0;
return virSecuritySELinuxRestoreFileLabel(mgr, savefile);
}
static int
virSecuritySELinuxVerify(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
return 0;
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
return -1;
}
if (secdef->type == VIR_DOMAIN_SECLABEL_STATIC) {
if (security_check_context(secdef->label) != 0) {
virReportError(VIR_ERR_XML_ERROR,
_("Invalid security label %s"), secdef->label);
return -1;
}
}
return 0;
}
static int
virSecuritySELinuxSetProcessLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def)
{
/* TODO: verify DOI */
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->label)
return 0;
VIR_DEBUG("label=%s", secdef->label);
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
if (security_getenforce() == 1)
return -1;
}
if (setexeccon_raw(secdef->label) == -1) {
virReportSystemError(errno,
_("unable to set security context '%s'"),
secdef->label);
if (security_getenforce() == 1)
return -1;
}
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return 0;
}
security: add new virSecurityManagerSetChildProcessLabel API The existing virSecurityManagerSetProcessLabel() API is designed so that it must be called after forking the child process, but before exec'ing the child. Due to the way the virCommand API works, that means it needs to be put in a "hook" function that virCommand is told to call out to at that time. Setting the child process label is a basic enough need when executing any process that virCommand should have a method of doing that. But virCommand must be told what label to set, and only the security driver knows the answer to that question. The new virSecurityManagerSet*Child*ProcessLabel() API is the way to transfer the knowledge about what label to set from the security driver to the virCommand object. It is given a virCommandPtr, and each security driver calls the appropriate virCommand* API to tell virCommand what to do between fork and exec. 1) in the case of the DAC security driver, it calls virCommandSetUID/GID() to set a uid and gid that must be set for the child process. 2) for the SELinux security driver, it calls virCommandSetSELinuxLabel() to save a copy of the char* that will be sent to setexeccon_raw() *after forking the child process*. 3) for the AppArmor security drivers, it calls virCommandSetAppArmorProfile() to save a copy of the char* that will be sent to aa_change_profile() *after forking the child process*. With this new API in place, we will be able to remove virSecurityManagerSetProcessLabel() from any virCommand pre-exec hooks. (Unfortunately, the LXC driver uses clone() rather than virCommand, so it can't take advantage of this new security driver API, meaning that we need to keep around the older virSecurityManagerSetProcessLabel(), at least for now.)
2013-02-01 20:02:03 +00:00
static int
virSecuritySELinuxSetChildProcessLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def,
virCommandPtr cmd)
security: add new virSecurityManagerSetChildProcessLabel API The existing virSecurityManagerSetProcessLabel() API is designed so that it must be called after forking the child process, but before exec'ing the child. Due to the way the virCommand API works, that means it needs to be put in a "hook" function that virCommand is told to call out to at that time. Setting the child process label is a basic enough need when executing any process that virCommand should have a method of doing that. But virCommand must be told what label to set, and only the security driver knows the answer to that question. The new virSecurityManagerSet*Child*ProcessLabel() API is the way to transfer the knowledge about what label to set from the security driver to the virCommand object. It is given a virCommandPtr, and each security driver calls the appropriate virCommand* API to tell virCommand what to do between fork and exec. 1) in the case of the DAC security driver, it calls virCommandSetUID/GID() to set a uid and gid that must be set for the child process. 2) for the SELinux security driver, it calls virCommandSetSELinuxLabel() to save a copy of the char* that will be sent to setexeccon_raw() *after forking the child process*. 3) for the AppArmor security drivers, it calls virCommandSetAppArmorProfile() to save a copy of the char* that will be sent to aa_change_profile() *after forking the child process*. With this new API in place, we will be able to remove virSecurityManagerSetProcessLabel() from any virCommand pre-exec hooks. (Unfortunately, the LXC driver uses clone() rather than virCommand, so it can't take advantage of this new security driver API, meaning that we need to keep around the older virSecurityManagerSetProcessLabel(), at least for now.)
2013-02-01 20:02:03 +00:00
{
/* TODO: verify DOI */
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->label)
security: add new virSecurityManagerSetChildProcessLabel API The existing virSecurityManagerSetProcessLabel() API is designed so that it must be called after forking the child process, but before exec'ing the child. Due to the way the virCommand API works, that means it needs to be put in a "hook" function that virCommand is told to call out to at that time. Setting the child process label is a basic enough need when executing any process that virCommand should have a method of doing that. But virCommand must be told what label to set, and only the security driver knows the answer to that question. The new virSecurityManagerSet*Child*ProcessLabel() API is the way to transfer the knowledge about what label to set from the security driver to the virCommand object. It is given a virCommandPtr, and each security driver calls the appropriate virCommand* API to tell virCommand what to do between fork and exec. 1) in the case of the DAC security driver, it calls virCommandSetUID/GID() to set a uid and gid that must be set for the child process. 2) for the SELinux security driver, it calls virCommandSetSELinuxLabel() to save a copy of the char* that will be sent to setexeccon_raw() *after forking the child process*. 3) for the AppArmor security drivers, it calls virCommandSetAppArmorProfile() to save a copy of the char* that will be sent to aa_change_profile() *after forking the child process*. With this new API in place, we will be able to remove virSecurityManagerSetProcessLabel() from any virCommand pre-exec hooks. (Unfortunately, the LXC driver uses clone() rather than virCommand, so it can't take advantage of this new security driver API, meaning that we need to keep around the older virSecurityManagerSetProcessLabel(), at least for now.)
2013-02-01 20:02:03 +00:00
return 0;
VIR_DEBUG("label=%s", secdef->label);
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
security: add new virSecurityManagerSetChildProcessLabel API The existing virSecurityManagerSetProcessLabel() API is designed so that it must be called after forking the child process, but before exec'ing the child. Due to the way the virCommand API works, that means it needs to be put in a "hook" function that virCommand is told to call out to at that time. Setting the child process label is a basic enough need when executing any process that virCommand should have a method of doing that. But virCommand must be told what label to set, and only the security driver knows the answer to that question. The new virSecurityManagerSet*Child*ProcessLabel() API is the way to transfer the knowledge about what label to set from the security driver to the virCommand object. It is given a virCommandPtr, and each security driver calls the appropriate virCommand* API to tell virCommand what to do between fork and exec. 1) in the case of the DAC security driver, it calls virCommandSetUID/GID() to set a uid and gid that must be set for the child process. 2) for the SELinux security driver, it calls virCommandSetSELinuxLabel() to save a copy of the char* that will be sent to setexeccon_raw() *after forking the child process*. 3) for the AppArmor security drivers, it calls virCommandSetAppArmorProfile() to save a copy of the char* that will be sent to aa_change_profile() *after forking the child process*. With this new API in place, we will be able to remove virSecurityManagerSetProcessLabel() from any virCommand pre-exec hooks. (Unfortunately, the LXC driver uses clone() rather than virCommand, so it can't take advantage of this new security driver API, meaning that we need to keep around the older virSecurityManagerSetProcessLabel(), at least for now.)
2013-02-01 20:02:03 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
if (security_getenforce() == 1)
return -1;
}
/* save in cmd to be set after fork/before child process is exec'ed */
virCommandSetSELinuxLabel(cmd, secdef->label);
return 0;
}
static int
virSecuritySELinuxSetDaemonSocketLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def)
{
/* TODO: verify DOI */
virSecurityLabelDefPtr secdef;
security_context_t scon = NULL;
char *str = NULL;
int rc = -1;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->label)
return 0;
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
goto done;
}
if (getcon_raw(&scon) == -1) {
virReportSystemError(errno,
_("unable to get current process context '%s'"),
secdef->label);
goto done;
}
if (!(str = virSecuritySELinuxContextAddRange(secdef->label, scon)))
goto done;
VIR_DEBUG("Setting VM %s socket context %s", def->name, str);
if (setsockcreatecon_raw(str) == -1) {
virReportSystemError(errno,
_("unable to set socket security context '%s'"), str);
goto done;
}
rc = 0;
done:
if (security_getenforce() != 1)
rc = 0;
freecon(scon);
VIR_FREE(str);
return rc;
}
static int
virSecuritySELinuxSetSocketLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr vm)
{
virSecurityLabelDefPtr secdef;
int rc = -1;
secdef = virDomainDefGetSecurityLabelDef(vm, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->label)
return 0;
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
goto done;
}
VIR_DEBUG("Setting VM %s socket context %s",
vm->name, secdef->label);
if (setsockcreatecon_raw(secdef->label) == -1) {
virReportSystemError(errno,
_("unable to set socket security context '%s'"),
secdef->label);
goto done;
}
rc = 0;
done:
if (security_getenforce() != 1)
rc = 0;
return rc;
}
static int
virSecuritySELinuxClearSocketLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def)
{
/* TODO: verify DOI */
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->label)
return 0;
if (STRNEQ(SECURITY_SELINUX_NAME, secdef->model)) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("security label driver mismatch: "
"'%s' model configured for domain, but "
"hypervisor driver is '%s'."),
secdef->model, SECURITY_SELINUX_NAME);
if (security_getenforce() == 1)
return -1;
}
if (setsockcreatecon_raw(NULL) == -1) {
virReportSystemError(errno,
_("unable to clear socket security context '%s'"),
secdef->label);
if (security_getenforce() == 1)
return -1;
}
return 0;
}
static int
virSecuritySELinuxSetSecurityChardevCallback(virDomainDefPtr def,
virDomainChrDefPtr dev ATTRIBUTE_UNUSED,
void *opaque)
{
struct _virSecuritySELinuxChardevCallbackData *data = opaque;
return virSecuritySELinuxSetChardevLabel(data->mgr, def, dev->source,
data->chardevStdioLogd);
}
static int
virSecuritySELinuxSetSecuritySmartcardCallback(virDomainDefPtr def,
virDomainSmartcardDefPtr dev,
void *opaque)
{
const char *database;
virSecurityManagerPtr mgr = opaque;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
switch (dev->type) {
case VIR_DOMAIN_SMARTCARD_TYPE_HOST:
break;
case VIR_DOMAIN_SMARTCARD_TYPE_HOST_CERTIFICATES:
database = dev->data.cert.database;
if (!database)
database = VIR_DOMAIN_SMARTCARD_DEFAULT_DATABASE;
return virSecuritySELinuxSetFilecon(mgr, database, data->content_context);
case VIR_DOMAIN_SMARTCARD_TYPE_PASSTHROUGH:
return virSecuritySELinuxSetChardevLabel(mgr, def,
dev->data.passthru, false);
default:
virReportError(VIR_ERR_INTERNAL_ERROR,
_("unknown smartcard type %d"),
dev->type);
return -1;
}
return 0;
}
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
static int
virSecuritySELinuxSetAllLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
const char *stdin_path,
bool chardevStdioLogd)
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
{
size_t i;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
if (!secdef || !secdef->relabel || data->skipAllLabel)
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return 0;
for (i = 0; i < def->ndisks; i++) {
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
/* XXX fixme - we need to recursively label the entire tree :-( */
conf: move host disk type to util/ A continuation of the migration of disk details to virstoragefile. This patch moves a single enum, but converting the name has quite a bit of fallout. * src/conf/domain_conf.h (virDomainDiskType): Move... * src/util/virstoragefile.h (virStorageType): ...and rename. * src/bhyve/bhyve_command.c (bhyveBuildDiskArgStr) (virBhyveProcessBuildLoadCmd): Update clients. * src/conf/domain_conf.c (virDomainDiskSourceDefParse) (virDomainDiskDefParseXML, virDomainDiskSourceDefFormatInternal) (virDomainDiskDefFormat, virDomainDiskGetActualType) (virDomainDiskDefForeachPath, virDomainDiskSourceIsBlockType): Likewise. * src/conf/snapshot_conf.h (_virDomainSnapshotDiskDef): Likewise. * src/conf/snapshot_conf.c (virDomainSnapshotDiskDefParseXML) (virDomainSnapshotAlignDisks, virDomainSnapshotDiskDefFormat): Likewise. * src/esx/esx_driver.c (esxAutodetectSCSIControllerModel) (esxDomainDefineXML): Likewise. * src/locking/domain_lock.c (virDomainLockManagerAddDisk): Likewise. * src/lxc/lxc_controller.c (virLXCControllerSetupLoopDeviceDisk) (virLXCControllerSetupNBDDeviceDisk) (virLXCControllerSetupLoopDevices, virLXCControllerSetupDisk): Likewise. * src/parallels/parallels_driver.c (parallelsGetHddInfo): Likewise. * src/phyp/phyp_driver.c (phypDiskType): Likewise. * src/qemu/qemu_command.c (qemuGetDriveSourceString) (qemuDomainDiskGetSourceString, qemuBuildDriveStr) (qemuBuildCommandLine, qemuParseCommandLineDisk) (qemuParseCommandLine): Likewise. * src/qemu/qemu_conf.c (qemuCheckSharedDevice) (qemuTranslateDiskSourcePool) (qemuTranslateSnapshotDiskSourcePool): Likewise. * src/qemu/qemu_domain.c (qemuDomainDeviceDefPostParse) (qemuDomainDetermineDiskChain): Likewise. * src/qemu/qemu_driver.c (qemuDomainGetBlockInfo) (qemuDomainSnapshotPrepareDiskExternalBackingInactive) (qemuDomainSnapshotPrepareDiskExternalBackingActive) (qemuDomainSnapshotPrepareDiskExternalOverlayActive) (qemuDomainSnapshotPrepareDiskExternalOverlayInactive) (qemuDomainSnapshotPrepareDiskInternal) (qemuDomainSnapshotPrepare) (qemuDomainSnapshotCreateSingleDiskActive): Likewise. * src/qemu/qemu_hotplug.c (qemuDomainChangeEjectableMedia): Likewise. * src/qemu/qemu_migration.c (qemuMigrationIsSafe): Likewise. * src/security/security_apparmor.c (AppArmorRestoreSecurityImageLabel) (AppArmorSetSecurityImageLabel): Likewise. * src/security/security_dac.c (virSecurityDACSetSecurityImageLabel) (virSecurityDACRestoreSecurityImageLabelInt) (virSecurityDACSetSecurityAllLabel): Likewise. * src/security/security_selinux.c (virSecuritySELinuxRestoreSecurityImageLabelInt) (virSecuritySELinuxSetSecurityImageLabel) (virSecuritySELinuxSetSecurityAllLabel): Likewise. * src/storage/storage_backend.c (virStorageFileBackendForType): Likewise. * src/storage/storage_backend_fs.c (virStorageFileBackendFile) (virStorageFileBackendBlock): Likewise. * src/storage/storage_backend_gluster.c (virStorageFileBackendGluster): Likewise. * src/vbox/vbox_tmpl.c (vboxDomainGetXMLDesc, vboxAttachDrives) (vboxDomainAttachDeviceImpl, vboxDomainDetachDevice): Likewise. * src/vmware/vmware_conf.c (vmwareVmxPath): Likewise. * src/vmx/vmx.c (virVMXParseDisk, virVMXFormatDisk) (virVMXFormatFloppy): Likewise. * src/xenxs/xen_sxpr.c (xenParseSxprDisks, xenParseSxpr) (xenFormatSxprDisk): Likewise. * src/xenxs/xen_xm.c (xenParseXM, xenFormatXMDisk): Likewise. * tests/securityselinuxlabeltest.c (testSELinuxLoadDef): Likewise. * src/libvirt_private.syms (domain_conf.h): Move symbols... (virstoragefile.h): ...as appropriate. Signed-off-by: Eric Blake <eblake@redhat.com>
2014-03-27 21:57:49 +00:00
if (virDomainDiskGetType(def->disks[i]) == VIR_STORAGE_TYPE_DIR) {
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
VIR_WARN("Unable to relabel directory tree %s for disk %s",
virDomainDiskGetSource(def->disks[i]),
def->disks[i]->dst);
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
continue;
}
if (virSecuritySELinuxSetDiskLabel(mgr,
def, def->disks[i]) < 0)
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return -1;
}
/* XXX fixme process def->fss if relabel == true */
for (i = 0; i < def->nhostdevs; i++) {
if (virSecuritySELinuxSetHostdevLabel(mgr,
def,
def->hostdevs[i],
NULL) < 0)
Refactor setup & cleanup of security labels in security driver The current security driver architecture has the following split of logic * domainGenSecurityLabel Allocate the unique label for the domain about to be started * domainGetSecurityLabel Retrieve the current live security label for a process * domainSetSecurityLabel Apply the previously allocated label to the current process Setup all disk image / device labelling * domainRestoreSecurityLabel Restore the original disk image / device labelling. Release the unique label for the domain The 'domainSetSecurityLabel' method is special because it runs in the context of the child process between the fork + exec. This is require in order to set the process label. It is not required in order to label disks/devices though. Having the disk labelling code run in the child process limits what it can do. In particularly libvirtd would like to remember the current disk image label, and only change shared image labels for the first VM to start. This requires use & update of global state in the libvirtd daemon, and thus cannot run in the child process context. The solution is to split domainSetSecurityLabel into two parts, one applies process label, and the other handles disk image labelling. At the same time domainRestoreSecurityLabel is similarly split, just so that it matches the style. Thus the previous 4 methods are replaced by the following 6 new methods * domainGenSecurityLabel Allocate the unique label for the domain about to be started No actual change here. * domainReleaseSecurityLabel Release the unique label for the domain * domainGetSecurityProcessLabel Retrieve the current live security label for a process Merely renamed for clarity. * domainSetSecurityProcessLabel Apply the previously allocated label to the current process * domainRestoreSecurityAllLabel Restore the original disk image / device labelling. * domainSetSecurityAllLabel Setup all disk image / device labelling The SELinux and AppArmour drivers are then updated to comply with this new spec. Notice that the AppArmour driver was actually a little different. It was creating its profile for the disk image and device labels in the 'domainGenSecurityLabel' method, where as the SELinux driver did it in 'domainSetSecurityLabel'. With the new method split, we can have consistency, with both drivers doing that in the domainSetSecurityAllLabel method. NB, the AppArmour changes here haven't been compiled so may not build.
2010-01-11 11:04:40 +00:00
return -1;
}
for (i = 0; i < def->ninputs; i++) {
if (virSecuritySELinuxSetInputLabel(mgr, def, def->inputs[i]) < 0)
return -1;
}
for (i = 0; i < def->nmems; i++) {
if (virSecuritySELinuxSetMemoryLabel(mgr, def, def->mems[i]) < 0)
return -1;
}
if (def->tpm) {
if (virSecuritySELinuxSetTPMFileLabel(mgr, def, def->tpm) < 0)
return -1;
}
struct _virSecuritySELinuxChardevCallbackData chardevData = {
.mgr = mgr,
.chardevStdioLogd = chardevStdioLogd
};
if (virDomainChrDefForeach(def,
true,
virSecuritySELinuxSetSecurityChardevCallback,
&chardevData) < 0)
return -1;
if (virDomainSmartcardDefForeach(def,
true,
virSecuritySELinuxSetSecuritySmartcardCallback,
mgr) < 0)
return -1;
/* This is different than kernel or initrd. The nvram store
* is really a disk, qemu can read and write to it. */
if (def->os.loader && def->os.loader->nvram &&
secdef && secdef->imagelabel &&
virSecuritySELinuxSetFilecon(mgr, def->os.loader->nvram,
secdef->imagelabel) < 0)
return -1;
if (def->os.kernel &&
virSecuritySELinuxSetFilecon(mgr, def->os.kernel,
data->content_context) < 0)
return -1;
if (def->os.initrd &&
virSecuritySELinuxSetFilecon(mgr, def->os.initrd,
data->content_context) < 0)
return -1;
2013-03-14 04:49:44 +00:00
if (def->os.dtb &&
virSecuritySELinuxSetFilecon(mgr, def->os.dtb,
data->content_context) < 0)
2013-03-14 04:49:44 +00:00
return -1;
if (def->os.slic_table &&
virSecuritySELinuxSetFilecon(mgr, def->os.slic_table,
data->content_context) < 0)
return -1;
if (stdin_path &&
virSecuritySELinuxSetFilecon(mgr, stdin_path,
data->content_context) < 0)
return -1;
return 0;
}
static int
virSecuritySELinuxSetImageFDLabel(virSecurityManagerPtr mgr ATTRIBUTE_UNUSED,
virDomainDefPtr def,
int fd)
{
virSecurityLabelDefPtr secdef;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!secdef || !secdef->imagelabel)
return 0;
return virSecuritySELinuxFSetFilecon(fd, secdef->imagelabel);
}
static int
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
virSecuritySELinuxSetTapFDLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
int fd)
{
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
struct stat buf;
security_context_t fcon = NULL;
virSecurityLabelDefPtr secdef;
char *str = NULL, *proc = NULL, *fd_path = NULL;
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
int rc = -1;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
if (!secdef || !secdef->label)
return 0;
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
if (fstat(fd, &buf) < 0) {
virReportSystemError(errno, _("cannot stat tap fd %d"), fd);
goto cleanup;
}
if ((buf.st_mode & S_IFMT) != S_IFCHR) {
virReportError(VIR_ERR_INTERNAL_ERROR,
_("tap fd %d is not character device"), fd);
goto cleanup;
}
/* Label /dev/tap.* devices only. Leave /dev/net/tun alone! */
if (virAsprintf(&proc, "/proc/self/fd/%d", fd) == -1)
goto cleanup;
if (virFileResolveLink(proc, &fd_path) < 0) {
virReportSystemError(errno,
_("Unable to resolve link: %s"), proc);
goto cleanup;
}
if (!STRPREFIX(fd_path, "/dev/tap")) {
VIR_DEBUG("fd=%d points to %s not setting SELinux label",
fd, fd_path);
rc = 0;
goto cleanup;
}
if (getContext(mgr, "/dev/tap*", buf.st_mode, &fcon) < 0) {
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
virReportError(VIR_ERR_INTERNAL_ERROR,
_("cannot lookup default selinux label for tap fd %d"), fd);
goto cleanup;
}
if (!(str = virSecuritySELinuxContextAddRange(secdef->label, fcon))) {
goto cleanup;
} else {
rc = virSecuritySELinuxFSetFilecon(fd, str);
}
cleanup:
freecon(fcon);
VIR_FREE(fd_path);
VIR_FREE(proc);
virSecuritySELinuxSetTapFDLabel: Temporarily revert to old behavior https://bugzilla.redhat.com/show_bug.cgi?id=1141879 A long time ago I've implemented support for so called multiqueue net. The idea was to let guest network traffic be processed by multiple host CPUs and thus increasing performance. However, this behavior is enabled by QEMU via special ioctl() iterated over the all tap FDs passed in by libvirt. Unfortunately, SELinux comes in and disallows the ioctl() call because the /dev/net/tun has label system_u:object_r:tun_tap_device_t:s0 and 'attach_queue' ioctl() is not allowed on tun_tap_device_t type. So after discussion with a SELinux developer we've decided that the FDs passed to the QEMU should be labelled with svirt_t type and SELinux policy will allow the ioctl(). Therefore I've made a patch (cf976d9dcf4e592261b14f03572) that does exactly this. The patch was fixed then by a4431931393aeb1ac5893f121151fa3df4fde612 and b635b7a1af0e64754016d758376f382470bc11e7. However, things are not that easy - even though the API to label FD is called (fsetfilecon_raw) the underlying file is labelled too! So effectively we are mangling /dev/net/tun label. Yes, that broke dozen of other application from openvpn, or boxes, to qemu running other domains. The best solution would be if SELinux provides a way to label an FD only, which could be then labeled when passed to the qemu. However that's a long path to go and we should fix this regression AQAP. So I went to talk to the SELinux developer again and we agreed on temporary solution that: 1) All the three patches are reverted 2) SELinux temporarily allows 'attach_queue' on the tun_tap_device_t Signed-off-by: Michal Privoznik <mprivozn@redhat.com>
2014-09-18 13:17:29 +00:00
VIR_FREE(str);
return rc;
}
static char *
virSecuritySELinuxGenImageLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def)
{
virSecurityLabelDefPtr secdef;
virSecuritySELinuxDataPtr data = virSecurityManagerGetPrivateData(mgr);
const char *range;
context_t ctx = NULL;
char *label = NULL;
char *mcs = NULL;
secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (secdef == NULL)
goto cleanup;
if (secdef->label) {
ctx = context_new(secdef->label);
if (!ctx) {
virReportSystemError(errno, _("unable to create selinux context for: %s"),
secdef->label);
goto cleanup;
}
range = context_range_get(ctx);
if (range) {
if (VIR_STRDUP(mcs, range) < 0)
goto cleanup;
if (!(label = virSecuritySELinuxGenNewContext(data->file_context,
mcs, true)))
goto cleanup;
}
}
cleanup:
context_free(ctx);
VIR_FREE(mcs);
return label;
}
static char *
virSecuritySELinuxGetSecurityMountOptions(virSecurityManagerPtr mgr,
virDomainDefPtr def)
{
char *opts = NULL;
virSecurityLabelDefPtr secdef;
if ((secdef = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME))) {
if (!secdef->imagelabel)
secdef->imagelabel = virSecuritySELinuxGenImageLabel(mgr, def);
if (secdef->imagelabel &&
virAsprintf(&opts,
",context=\"%s\"",
(const char*) secdef->imagelabel) < 0)
return NULL;
}
if (!opts && VIR_STRDUP(opts, "") < 0)
return NULL;
VIR_DEBUG("imageLabel=%s opts=%s",
secdef ? secdef->imagelabel : "(null)", opts);
return opts;
}
static int
virSecuritySELinuxDomainSetPathLabel(virSecurityManagerPtr mgr,
virDomainDefPtr def,
const char *path)
{
virSecurityLabelDefPtr seclabel;
seclabel = virDomainDefGetSecurityLabelDef(def, SECURITY_SELINUX_NAME);
if (!seclabel || !seclabel->relabel)
return 0;
return virSecuritySELinuxSetFilecon(mgr, path, seclabel->imagelabel);
}
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
virSecurityDriver virSecurityDriverSELinux = {
.privateDataLen = sizeof(virSecuritySELinuxData),
.name = SECURITY_SELINUX_NAME,
.probe = virSecuritySELinuxDriverProbe,
.open = virSecuritySELinuxDriverOpen,
.close = virSecuritySELinuxDriverClose,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.getModel = virSecuritySELinuxGetModel,
.getDOI = virSecuritySELinuxGetDOI,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.transactionStart = virSecuritySELinuxTransactionStart,
.transactionCommit = virSecuritySELinuxTransactionCommit,
.transactionAbort = virSecuritySELinuxTransactionAbort,
.domainSecurityVerify = virSecuritySELinuxVerify,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainSetSecurityDiskLabel = virSecuritySELinuxSetDiskLabel,
.domainRestoreSecurityDiskLabel = virSecuritySELinuxRestoreDiskLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainSetSecurityImageLabel = virSecuritySELinuxSetImageLabel,
.domainRestoreSecurityImageLabel = virSecuritySELinuxRestoreImageLabel,
.domainSetSecurityMemoryLabel = virSecuritySELinuxSetMemoryLabel,
.domainRestoreSecurityMemoryLabel = virSecuritySELinuxRestoreMemoryLabel,
.domainSetSecurityDaemonSocketLabel = virSecuritySELinuxSetDaemonSocketLabel,
.domainSetSecuritySocketLabel = virSecuritySELinuxSetSocketLabel,
.domainClearSecuritySocketLabel = virSecuritySELinuxClearSocketLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainGenSecurityLabel = virSecuritySELinuxGenLabel,
.domainReserveSecurityLabel = virSecuritySELinuxReserveLabel,
.domainReleaseSecurityLabel = virSecuritySELinuxReleaseLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainGetSecurityProcessLabel = virSecuritySELinuxGetProcessLabel,
.domainSetSecurityProcessLabel = virSecuritySELinuxSetProcessLabel,
.domainSetSecurityChildProcessLabel = virSecuritySELinuxSetChildProcessLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainSetSecurityAllLabel = virSecuritySELinuxSetAllLabel,
.domainRestoreSecurityAllLabel = virSecuritySELinuxRestoreAllLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainSetSecurityHostdevLabel = virSecuritySELinuxSetHostdevLabel,
.domainRestoreSecurityHostdevLabel = virSecuritySELinuxRestoreHostdevLabel,
Refactor the security drivers to simplify usage The current security driver usage requires horrible code like if (driver->securityDriver && driver->securityDriver->domainSetSecurityHostdevLabel && driver->securityDriver->domainSetSecurityHostdevLabel(driver->securityDriver, vm, hostdev) < 0) This pair of checks for NULL clutters up the code, making the driver calls 2 lines longer than they really need to be. The goal of the patchset is to change the calling convention to simply if (virSecurityManagerSetHostdevLabel(driver->securityDriver, vm, hostdev) < 0) The first check for 'driver->securityDriver' being NULL is removed by introducing a 'no op' security driver that will always be present if no real driver is enabled. This guarentees driver->securityDriver != NULL. The second check for 'driver->securityDriver->domainSetSecurityHostdevLabel' being non-NULL is hidden in a new abstraction called virSecurityManager. This separates the driver callbacks, from main internal API. The addition of a virSecurityManager object, that is separate from the virSecurityDriver struct also allows for security drivers to carry state / configuration information directly. Thus the DAC/Stack drivers from src/qemu which used to pull config from 'struct qemud_driver' can now be moved into the 'src/security' directory and store their config directly. * src/qemu/qemu_conf.h, src/qemu/qemu_driver.c: Update to use new virSecurityManager APIs * src/qemu/qemu_security_dac.c, src/qemu/qemu_security_dac.h src/qemu/qemu_security_stacked.c, src/qemu/qemu_security_stacked.h: Move into src/security directory * src/security/security_stack.c, src/security/security_stack.h, src/security/security_dac.c, src/security/security_dac.h: Generic versions of previous QEMU specific drivers * src/security/security_apparmor.c, src/security/security_apparmor.h, src/security/security_driver.c, src/security/security_driver.h, src/security/security_selinux.c, src/security/security_selinux.h: Update to take virSecurityManagerPtr object as the first param in all callbacks * src/security/security_nop.c, src/security/security_nop.h: Stub implementation of all security driver APIs. * src/security/security_manager.h, src/security/security_manager.c: New internal API for invoking security drivers * src/libvirt.c: Add missing debug for security APIs
2010-11-17 20:26:30 +00:00
.domainSetSavedStateLabel = virSecuritySELinuxSetSavedStateLabel,
.domainRestoreSavedStateLabel = virSecuritySELinuxRestoreSavedStateLabel,
.domainSetSecurityImageFDLabel = virSecuritySELinuxSetImageFDLabel,
.domainSetSecurityTapFDLabel = virSecuritySELinuxSetTapFDLabel,
.domainGetSecurityMountOptions = virSecuritySELinuxGetSecurityMountOptions,
.getBaseLabel = virSecuritySELinuxGetBaseLabel,
.domainSetPathLabel = virSecuritySELinuxDomainSetPathLabel,
};