libvirt/src/storage_backend_fs.c

1281 lines
41 KiB
C

/*
* storage_backend_fs.c: storage backend for FS and directory handling
*
* Copyright (C) 2007-2009 Red Hat, Inc.
* Copyright (C) 2007-2008 Daniel P. Berrange
*
* 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, write to the Free Software
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
*
* Author: Daniel P. Berrange <berrange@redhat.com>
*/
#include <config.h>
#include <sys/statvfs.h>
#include <sys/types.h>
#include <sys/stat.h>
#include <stdio.h>
#include <dirent.h>
#include <errno.h>
#include <fcntl.h>
#include <unistd.h>
#include <string.h>
#include <libxml/parser.h>
#include <libxml/tree.h>
#include <libxml/xpath.h>
#include "virterror_internal.h"
#include "storage_backend_fs.h"
#include "storage_conf.h"
#include "util.h"
#include "memory.h"
#include "xml.h"
enum lv_endian {
LV_LITTLE_ENDIAN = 1, /* 1234 */
LV_BIG_ENDIAN /* 4321 */
};
enum {
BACKING_STORE_OK,
BACKING_STORE_INVALID,
BACKING_STORE_ERROR,
};
static int cowGetBackingStore(virConnectPtr, char **,
const unsigned char *, size_t);
static int qcowXGetBackingStore(virConnectPtr, char **,
const unsigned char *, size_t);
static int vmdk4GetBackingStore(virConnectPtr, char **,
const unsigned char *, size_t);
/* Either 'magic' or 'extension' *must* be provided */
struct FileTypeInfo {
int type; /* One of the constants above */
const char *magic; /* Optional string of file magic
* to check at head of file */
const char *extension; /* Optional file extension to check */
enum lv_endian endian; /* Endianness of file format */
int versionOffset; /* Byte offset from start of file
* where we find version number,
* -1 to skip version test */
int versionNumber; /* Version number to validate */
int sizeOffset; /* Byte offset from start of file
* where we find capacity info,
* -1 to use st_size as capacity */
int sizeBytes; /* Number of bytes for size field */
int sizeMultiplier; /* A scaling factor if size is not in bytes */
/* Store a COW base image path (possibly relative),
* or NULL if there is no COW base image, to RES;
* return BACKING_STORE_* */
int (*getBackingStore)(virConnectPtr conn, char **res,
const unsigned char *buf, size_t buf_size);
};
struct FileTypeInfo const fileTypeInfo[] = {
/* Bochs */
/* XXX Untested
{ VIR_STORAGE_VOL_BOCHS, "Bochs Virtual HD Image", NULL,
LV_LITTLE_ENDIAN, 64, 0x20000,
32+16+16+4+4+4+4+4, 8, 1, NULL },*/
/* CLoop */
/* XXX Untested
{ VIR_STORAGE_VOL_CLOOP, "#!/bin/sh\n#V2.0 Format\nmodprobe cloop file=$0 && mount -r -t iso9660 /dev/cloop $1\n", NULL,
LV_LITTLE_ENDIAN, -1, 0,
-1, 0, 0, NULL }, */
/* Cow */
{ VIR_STORAGE_VOL_FILE_COW, "OOOM", NULL,
LV_BIG_ENDIAN, 4, 2,
4+4+1024+4, 8, 1, cowGetBackingStore },
/* DMG */
/* XXX QEMU says there's no magic for dmg, but we should check... */
{ VIR_STORAGE_VOL_FILE_DMG, NULL, ".dmg",
0, -1, 0,
-1, 0, 0, NULL },
/* XXX there's probably some magic for iso we can validate too... */
{ VIR_STORAGE_VOL_FILE_ISO, NULL, ".iso",
0, -1, 0,
-1, 0, 0, NULL },
/* Parallels */
/* XXX Untested
{ VIR_STORAGE_VOL_FILE_PARALLELS, "WithoutFreeSpace", NULL,
LV_LITTLE_ENDIAN, 16, 2,
16+4+4+4+4, 4, 512, NULL },
*/
/* QCow */
{ VIR_STORAGE_VOL_FILE_QCOW, "QFI", NULL,
LV_BIG_ENDIAN, 4, 1,
4+4+8+4+4, 8, 1, qcowXGetBackingStore },
/* QCow 2 */
{ VIR_STORAGE_VOL_FILE_QCOW2, "QFI", NULL,
LV_BIG_ENDIAN, 4, 2,
4+4+8+4+4, 8, 1, qcowXGetBackingStore },
/* VMDK 3 */
/* XXX Untested
{ VIR_STORAGE_VOL_FILE_VMDK, "COWD", NULL,
LV_LITTLE_ENDIAN, 4, 1,
4+4+4, 4, 512, NULL },
*/
/* VMDK 4 */
{ VIR_STORAGE_VOL_FILE_VMDK, "KDMV", NULL,
LV_LITTLE_ENDIAN, 4, 1,
4+4+4, 8, 512, vmdk4GetBackingStore },
/* Connectix / VirtualPC */
/* XXX Untested
{ VIR_STORAGE_VOL_FILE_VPC, "conectix", NULL,
LV_BIG_ENDIAN, -1, 0,
-1, 0, 0, NULL},
*/
};
#define VIR_FROM_THIS VIR_FROM_STORAGE
static int
cowGetBackingStore(virConnectPtr conn,
char **res,
const unsigned char *buf,
size_t buf_size)
{
#define COW_FILENAME_MAXLEN 1024
*res = NULL;
if (buf_size < 4+4+ COW_FILENAME_MAXLEN)
return BACKING_STORE_INVALID;
if (buf[4+4] == '\0') /* cow_header_v2.backing_file[0] */
return BACKING_STORE_OK;
*res = strndup ((const char*)buf + 4+4, COW_FILENAME_MAXLEN);
if (*res == NULL) {
virReportOOMError(conn);
return BACKING_STORE_ERROR;
}
return BACKING_STORE_OK;
}
static int
qcowXGetBackingStore(virConnectPtr conn,
char **res,
const unsigned char *buf,
size_t buf_size)
{
unsigned long long offset;
unsigned long size;
*res = NULL;
if (buf_size < 4+4+8+4)
return BACKING_STORE_INVALID;
offset = (((unsigned long long)buf[4+4] << 56)
| ((unsigned long long)buf[4+4+1] << 48)
| ((unsigned long long)buf[4+4+2] << 40)
| ((unsigned long long)buf[4+4+3] << 32)
| ((unsigned long long)buf[4+4+4] << 24)
| ((unsigned long long)buf[4+4+5] << 16)
| ((unsigned long long)buf[4+4+6] << 8)
| buf[4+4+7]); /* QCowHeader.backing_file_offset */
if (offset > buf_size)
return BACKING_STORE_INVALID;
size = ((buf[4+4+8] << 24)
| (buf[4+4+8+1] << 16)
| (buf[4+4+8+2] << 8)
| buf[4+4+8+3]); /* QCowHeader.backing_file_size */
if (size == 0)
return BACKING_STORE_OK;
if (offset + size > buf_size || offset + size < offset)
return BACKING_STORE_INVALID;
if (size + 1 == 0)
return BACKING_STORE_INVALID;
if (VIR_ALLOC_N(*res, size + 1) < 0) {
virReportOOMError(conn);
return BACKING_STORE_ERROR;
}
memcpy(*res, buf + offset, size);
(*res)[size] = '\0';
return BACKING_STORE_OK;
}
static int
vmdk4GetBackingStore(virConnectPtr conn,
char **res,
const unsigned char *buf,
size_t buf_size)
{
static const char prefix[] = "parentFileNameHint=\"";
char desc[20*512 + 1], *start, *end;
size_t len;
*res = NULL;
if (buf_size <= 0x200)
return BACKING_STORE_INVALID;
len = buf_size - 0x200;
if (len > sizeof(desc) - 1)
len = sizeof(desc) - 1;
memcpy(desc, buf + 0x200, len);
desc[len] = '\0';
start = strstr(desc, prefix);
if (start == NULL)
return BACKING_STORE_OK;
start += strlen(prefix);
end = strchr(start, '"');
if (end == NULL)
return BACKING_STORE_INVALID;
if (end == start)
return BACKING_STORE_OK;
*end = '\0';
*res = strdup(start);
if (*res == NULL) {
virReportOOMError(conn);
return BACKING_STORE_ERROR;
}
return BACKING_STORE_OK;
}
/**
* Return an absolute path corresponding to PATH, which is absolute or relative
* to the directory containing BASE_FILE, or NULL on error
*/
static char *absolutePathFromBaseFile(const char *base_file, const char *path)
{
size_t base_size, path_size;
char *res, *p;
if (*path == '/')
return strdup(path);
base_size = strlen(base_file) + 1;
path_size = strlen(path) + 1;
if (VIR_ALLOC_N(res, base_size - 1 + path_size) < 0)
return NULL;
memcpy(res, base_file, base_size);
p = strrchr(res, '/');
if (p != NULL)
p++;
else
p = res;
memcpy(p, path, path_size);
if (VIR_REALLOC_N(res, (p + path_size) - res) < 0) {
/* Ignore failure */
}
return res;
}
/**
* Probe the header of a file to determine what type of disk image
* it is, and info about its capacity if available.
*/
static int virStorageBackendProbeTarget(virConnectPtr conn,
virStorageVolTargetPtr target,
char **backingStore,
unsigned long long *allocation,
unsigned long long *capacity) {
int fd;
unsigned char head[20*512]; /* vmdk4GetBackingStore needs this much. */
int len, i, ret;
if (backingStore)
*backingStore = NULL;
if ((fd = open(target->path, O_RDONLY)) < 0) {
virReportSystemError(conn, errno,
_("cannot open volume '%s'"),
target->path);
return -1;
}
if ((ret = virStorageBackendUpdateVolTargetInfoFD(conn, target, fd,
allocation,
capacity)) < 0) {
close(fd);
return ret; /* Take care to propagate ret, it is not always -1 */
}
if ((len = read(fd, head, sizeof(head))) < 0) {
virReportSystemError(conn, errno,
_("cannot read header '%s'"),
target->path);
close(fd);
return -1;
}
close(fd);
/* First check file magic */
for (i = 0 ; i < ARRAY_CARDINALITY(fileTypeInfo) ; i++) {
int mlen;
if (fileTypeInfo[i].magic == NULL)
continue;
/* Validate magic data */
mlen = strlen(fileTypeInfo[i].magic);
if (mlen > len)
continue;
if (memcmp(head, fileTypeInfo[i].magic, mlen) != 0)
continue;
/* Validate version number info */
if (fileTypeInfo[i].versionNumber != -1) {
int version;
if (fileTypeInfo[i].endian == LV_LITTLE_ENDIAN) {
version = (head[fileTypeInfo[i].versionOffset+3] << 24) |
(head[fileTypeInfo[i].versionOffset+2] << 16) |
(head[fileTypeInfo[i].versionOffset+1] << 8) |
head[fileTypeInfo[i].versionOffset];
} else {
version = (head[fileTypeInfo[i].versionOffset] << 24) |
(head[fileTypeInfo[i].versionOffset+1] << 16) |
(head[fileTypeInfo[i].versionOffset+2] << 8) |
head[fileTypeInfo[i].versionOffset+3];
}
if (version != fileTypeInfo[i].versionNumber)
continue;
}
/* Optionally extract capacity from file */
if (fileTypeInfo[i].sizeOffset != -1 && capacity) {
if (fileTypeInfo[i].endian == LV_LITTLE_ENDIAN) {
*capacity =
((unsigned long long)head[fileTypeInfo[i].sizeOffset+7] << 56) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+6] << 48) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+5] << 40) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+4] << 32) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+3] << 24) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+2] << 16) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+1] << 8) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset]);
} else {
*capacity =
((unsigned long long)head[fileTypeInfo[i].sizeOffset] << 56) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+1] << 48) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+2] << 40) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+3] << 32) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+4] << 24) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+5] << 16) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+6] << 8) |
((unsigned long long)head[fileTypeInfo[i].sizeOffset+7]);
}
/* Avoid unlikely, but theoretically possible overflow */
if (*capacity > (ULLONG_MAX / fileTypeInfo[i].sizeMultiplier))
continue;
*capacity *= fileTypeInfo[i].sizeMultiplier;
}
/* Validation passed, we know the file format now */
target->format = fileTypeInfo[i].type;
if (fileTypeInfo[i].getBackingStore != NULL && backingStore) {
char *base;
switch (fileTypeInfo[i].getBackingStore(conn, &base, head, len)) {
case BACKING_STORE_OK:
break;
case BACKING_STORE_INVALID:
continue;
case BACKING_STORE_ERROR:
return -1;
}
if (base != NULL) {
*backingStore
= absolutePathFromBaseFile(target->path, base);
VIR_FREE(base);
if (*backingStore == NULL) {
virReportOOMError(conn);
return -1;
}
}
}
return 0;
}
/* No magic, so check file extension */
for (i = 0 ; i < ARRAY_CARDINALITY(fileTypeInfo) ; i++) {
if (fileTypeInfo[i].extension == NULL)
continue;
if (!virFileHasSuffix(target->path, fileTypeInfo[i].extension))
continue;
target->format = fileTypeInfo[i].type;
return 0;
}
/* All fails, so call it a raw file */
target->format = VIR_STORAGE_VOL_FILE_RAW;
return 0;
}
#if WITH_STORAGE_FS
#include <mntent.h>
struct _virNetfsDiscoverState {
const char *host;
virStoragePoolSourceList list;
};
typedef struct _virNetfsDiscoverState virNetfsDiscoverState;
static int
virStorageBackendFileSystemNetFindPoolSourcesFunc(virConnectPtr conn ATTRIBUTE_UNUSED,
virStoragePoolObjPtr pool ATTRIBUTE_UNUSED,
char **const groups,
void *data)
{
virNetfsDiscoverState *state = data;
const char *name, *path;
virStoragePoolSource *src;
path = groups[0];
name = strrchr(path, '/');
if (name == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
_("invalid netfs path (no /): %s"), path);
return -1;
}
name += 1;
if (*name == '\0') {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
_("invalid netfs path (ends in /): %s"), path);
return -1;
}
if (VIR_REALLOC_N(state->list.sources, state->list.nsources+1) < 0) {
virReportOOMError(conn);
return -1;
}
memset(state->list.sources + state->list.nsources, 0, sizeof(*state->list.sources));
src = state->list.sources + state->list.nsources++;
if (!(src->host.name = strdup(state->host)) ||
!(src->dir = strdup(path)))
return -1;
src->format = VIR_STORAGE_POOL_NETFS_NFS;
return 0;
}
static char *
virStorageBackendFileSystemNetFindPoolSources(virConnectPtr conn,
const char *srcSpec,
unsigned int flags ATTRIBUTE_UNUSED)
{
/*
* # showmount --no-headers -e HOSTNAME
* /tmp *
* /A dir demo1.foo.bar,demo2.foo.bar
*
* Extract directory name (including possible interior spaces ...).
*/
const char *regexes[] = {
"^(/.*\\S) +\\S+$"
};
int vars[] = {
1
};
xmlDocPtr doc = NULL;
xmlXPathContextPtr xpath_ctxt = NULL;
virNetfsDiscoverState state = {
.host = NULL,
.list = {
.type = VIR_STORAGE_POOL_NETFS,
.nsources = 0,
.sources = NULL
}
};
const char *prog[] = { SHOWMOUNT, "--no-headers", "--exports", NULL, NULL };
int exitstatus;
char *retval = NULL;
unsigned int i;
doc = xmlReadDoc((const xmlChar *)srcSpec, "srcSpec.xml", NULL,
XML_PARSE_NOENT | XML_PARSE_NONET |
XML_PARSE_NOERROR | XML_PARSE_NOWARNING);
if (doc == NULL) {
virStorageReportError(conn, VIR_ERR_XML_ERROR, "%s", _("bad <source> spec"));
goto cleanup;
}
xpath_ctxt = xmlXPathNewContext(doc);
if (xpath_ctxt == NULL) {
virReportOOMError(conn);
goto cleanup;
}
state.host = virXPathString(conn, "string(/source/host/@name)", xpath_ctxt);
if (!state.host || !state.host[0]) {
virStorageReportError(conn, VIR_ERR_XML_ERROR, "%s",
_("missing <host> in <source> spec"));
goto cleanup;
}
prog[3] = state.host;
if (virStorageBackendRunProgRegex(conn, NULL, prog, 1, regexes, vars,
virStorageBackendFileSystemNetFindPoolSourcesFunc,
&state, &exitstatus) < 0)
goto cleanup;
retval = virStoragePoolSourceListFormat(conn, &state.list);
if (retval == NULL) {
virReportOOMError(conn);
goto cleanup;
}
cleanup:
for (i = 0; i < state.list.nsources; i++)
virStoragePoolSourceFree(&state.list.sources[i]);
VIR_FREE(state.list.sources);
VIR_FREE(state.host);
xmlFreeDoc(doc);
xmlXPathFreeContext(xpath_ctxt);
return retval;
}
/**
* @conn connection to report errors against
* @pool storage pool to check for status
*
* Determine if a storage pool is already mounted
*
* Return 0 if not mounted, 1 if mounted, -1 on error
*/
static int
virStorageBackendFileSystemIsMounted(virConnectPtr conn,
virStoragePoolObjPtr pool) {
FILE *mtab;
struct mntent ent;
char buf[1024];
if ((mtab = fopen(_PATH_MOUNTED, "r")) == NULL) {
virReportSystemError(conn, errno,
_("cannot read mount list '%s'"),
_PATH_MOUNTED);
return -1;
}
while ((getmntent_r(mtab, &ent, buf, sizeof(buf))) != NULL) {
if (STREQ(ent.mnt_dir, pool->def->target.path)) {
fclose(mtab);
return 1;
}
}
fclose(mtab);
return 0;
}
/**
* @conn connection to report errors against
* @pool storage pool to mount
*
* Ensure that a FS storage pool is mounted on its target location.
* If already mounted, this is a no-op
*
* Returns 0 if successfully mounted, -1 on error
*/
static int
virStorageBackendFileSystemMount(virConnectPtr conn,
virStoragePoolObjPtr pool) {
char *src;
const char **mntargv;
/* 'mount -t auto' doesn't seem to auto determine nfs (or cifs),
* while plain 'mount' does. We have to craft separate argvs to
* accommodate this */
int netauto = (pool->def->type == VIR_STORAGE_POOL_NETFS &&
pool->def->source.format == VIR_STORAGE_POOL_NETFS_AUTO);
int source_index;
const char *netfs_auto_argv[] = {
MOUNT,
NULL, /* source path */
pool->def->target.path,
NULL,
};
const char *fs_argv[] = {
MOUNT,
"-t",
pool->def->type == VIR_STORAGE_POOL_FS ?
virStoragePoolFormatFileSystemTypeToString(pool->def->source.format) :
virStoragePoolFormatFileSystemNetTypeToString(pool->def->source.format),
NULL, /* Fill in shortly - careful not to add extra fields
before this */
pool->def->target.path,
NULL,
};
if (netauto) {
mntargv = netfs_auto_argv;
source_index = 1;
} else {
mntargv = fs_argv;
source_index = 3;
}
int ret;
if (pool->def->type == VIR_STORAGE_POOL_NETFS) {
if (pool->def->source.host.name == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source host"));
return -1;
}
if (pool->def->source.dir == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source path"));
return -1;
}
} else {
if (pool->def->source.ndevice != 1) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source device"));
return -1;
}
}
/* Short-circuit if already mounted */
if ((ret = virStorageBackendFileSystemIsMounted(conn, pool)) != 0) {
if (ret < 0)
return -1;
else
return 0;
}
if (pool->def->type == VIR_STORAGE_POOL_NETFS) {
if (VIR_ALLOC_N(src, strlen(pool->def->source.host.name) +
1 + strlen(pool->def->source.dir) + 1) < 0) {
virReportOOMError(conn);
return -1;
}
strcpy(src, pool->def->source.host.name);
strcat(src, ":");
strcat(src, pool->def->source.dir);
} else {
if ((src = strdup(pool->def->source.devices[0].path)) == NULL) {
virReportOOMError(conn);
return -1;
}
}
mntargv[source_index] = src;
if (virRun(conn, mntargv, NULL) < 0) {
VIR_FREE(src);
return -1;
}
VIR_FREE(src);
return 0;
}
/**
* @conn connection to report errors against
* @pool storage pool to unmount
*
* Ensure that a FS storage pool is not mounted on its target location.
* If already unmounted, this is a no-op
*
* Returns 0 if successfully unmounted, -1 on error
*/
static int
virStorageBackendFileSystemUnmount(virConnectPtr conn,
virStoragePoolObjPtr pool) {
const char *mntargv[3];
int ret;
if (pool->def->type == VIR_STORAGE_POOL_NETFS) {
if (pool->def->source.host.name == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source host"));
return -1;
}
if (pool->def->source.dir == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source dir"));
return -1;
}
} else {
if (pool->def->source.ndevice != 1) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("missing source device"));
return -1;
}
}
/* Short-circuit if already unmounted */
if ((ret = virStorageBackendFileSystemIsMounted(conn, pool)) != 1) {
if (ret < 0)
return -1;
else
return 0;
}
mntargv[0] = UMOUNT;
mntargv[1] = pool->def->target.path;
mntargv[2] = NULL;
if (virRun(conn, mntargv, NULL) < 0) {
return -1;
}
return 0;
}
#endif /* WITH_STORAGE_FS */
/**
* @conn connection to report errors against
* @pool storage pool to start
*
* Starts a directory or FS based storage pool.
*
* - If it is a FS based pool, mounts the unlying source device on the pool
*
* Returns 0 on success, -1 on error
*/
#if WITH_STORAGE_FS
static int
virStorageBackendFileSystemStart(virConnectPtr conn,
virStoragePoolObjPtr pool)
{
if (pool->def->type != VIR_STORAGE_POOL_DIR &&
virStorageBackendFileSystemMount(conn, pool) < 0)
return -1;
return 0;
}
#endif /* WITH_STORAGE_FS */
/**
* @conn connection to report errors against
* @pool storage pool to build
*
* Build a directory or FS based storage pool.
*
* - If it is a FS based pool, mounts the unlying source device on the pool
*
* Returns 0 on success, -1 on error
*/
static int
virStorageBackendFileSystemBuild(virConnectPtr conn,
virStoragePoolObjPtr pool,
unsigned int flags ATTRIBUTE_UNUSED)
{
int err;
if ((err = virFileMakePath(pool->def->target.path)) < 0) {
virReportSystemError(conn, err,
_("cannot create path '%s'"),
pool->def->target.path);
return -1;
}
return 0;
}
/**
* Iterate over the pool's directory and enumerate all disk images
* within it. This is non-recursive.
*/
static int
virStorageBackendFileSystemRefresh(virConnectPtr conn,
virStoragePoolObjPtr pool)
{
DIR *dir;
struct dirent *ent;
struct statvfs sb;
virStorageVolDefPtr vol = NULL;
if (!(dir = opendir(pool->def->target.path))) {
virReportSystemError(conn, errno,
_("cannot open path '%s'"),
pool->def->target.path);
goto cleanup;
}
while ((ent = readdir(dir)) != NULL) {
int ret;
char *backingStore;
if (VIR_ALLOC(vol) < 0)
goto no_memory;
if ((vol->name = strdup(ent->d_name)) == NULL)
goto no_memory;
vol->type = VIR_STORAGE_VOL_FILE;
vol->target.format = VIR_STORAGE_VOL_FILE_RAW; /* Real value is filled in during probe */
if (VIR_ALLOC_N(vol->target.path, strlen(pool->def->target.path) +
1 + strlen(vol->name) + 1) < 0)
goto no_memory;
strcpy(vol->target.path, pool->def->target.path);
strcat(vol->target.path, "/");
strcat(vol->target.path, vol->name);
if ((vol->key = strdup(vol->target.path)) == NULL)
goto no_memory;
if ((ret = virStorageBackendProbeTarget(conn,
&vol->target,
&backingStore,
&vol->allocation,
&vol->capacity) < 0)) {
if (ret == -1)
goto no_memory;
else {
/* Silently ignore non-regular files,
* eg '.' '..', 'lost+found' */
virStorageVolDefFree(vol);
vol = NULL;
continue;
}
}
if (backingStore != NULL) {
if (vol->target.format == VIR_STORAGE_VOL_FILE_QCOW2 &&
STRPREFIX("fmt:", backingStore)) {
char *fmtstr = backingStore + 4;
char *path = strchr(fmtstr, ':');
if (!path) {
VIR_FREE(backingStore);
} else {
*path = '\0';
if ((vol->backingStore.format =
virStorageVolFormatFileSystemTypeFromString(fmtstr)) < 0) {
VIR_FREE(backingStore);
} else {
memmove(backingStore, path, strlen(path) + 1);
vol->backingStore.path = backingStore;
if (virStorageBackendUpdateVolTargetInfo(conn,
&vol->backingStore,
NULL,
NULL) < 0)
VIR_FREE(vol->backingStore);
}
}
} else {
vol->backingStore.path = backingStore;
if ((ret = virStorageBackendProbeTarget(conn,
&vol->backingStore,
NULL, NULL, NULL)) < 0) {
if (ret == -1)
goto no_memory;
else {
/* Silently ignore non-regular files,
* eg '.' '..', 'lost+found' */
VIR_FREE(vol->backingStore);
}
}
}
}
if (VIR_REALLOC_N(pool->volumes.objs,
pool->volumes.count+1) < 0)
goto no_memory;
pool->volumes.objs[pool->volumes.count++] = vol;
vol = NULL;
}
closedir(dir);
if (statvfs(pool->def->target.path, &sb) < 0) {
virReportSystemError(conn, errno,
_("cannot statvfs path '%s'"),
pool->def->target.path);
return -1;
}
pool->def->capacity = ((unsigned long long)sb.f_frsize *
(unsigned long long)sb.f_blocks);
pool->def->available = ((unsigned long long)sb.f_bfree *
(unsigned long long)sb.f_bsize);
pool->def->allocation = pool->def->capacity - pool->def->available;
return 0;
no_memory:
virReportOOMError(conn);
/* fallthrough */
cleanup:
closedir(dir);
virStorageVolDefFree(vol);
virStoragePoolObjClearVols(pool);
return -1;
}
/**
* @conn connection to report errors against
* @pool storage pool to start
*
* Stops a directory or FS based storage pool.
*
* - If it is a FS based pool, unmounts the unlying source device on the pool
* - Releases all cached data about volumes
*/
#if WITH_STORAGE_FS
static int
virStorageBackendFileSystemStop(virConnectPtr conn,
virStoragePoolObjPtr pool)
{
if (pool->def->type != VIR_STORAGE_POOL_DIR &&
virStorageBackendFileSystemUnmount(conn, pool) < 0)
return -1;
return 0;
}
#endif /* WITH_STORAGE_FS */
/**
* @conn connection to report errors against
* @pool storage pool to build
*
* Build a directory or FS based storage pool.
*
* - If it is a FS based pool, mounts the unlying source device on the pool
*
* Returns 0 on success, -1 on error
*/
static int
virStorageBackendFileSystemDelete(virConnectPtr conn,
virStoragePoolObjPtr pool,
unsigned int flags ATTRIBUTE_UNUSED)
{
/* XXX delete all vols first ? */
if (unlink(pool->def->target.path) < 0) {
virReportSystemError(conn, errno,
_("cannot unlink path '%s'"),
pool->def->target.path);
return -1;
}
return 0;
}
/**
* Allocate a new file as a volume. This is either done directly
* for raw/sparse files, or by calling qemu-img/qcow-create for
* special kinds of files
*/
static int
virStorageBackendFileSystemVolCreate(virConnectPtr conn,
virStoragePoolObjPtr pool,
virStorageVolDefPtr vol)
{
int fd;
if (VIR_ALLOC_N(vol->target.path, strlen(pool->def->target.path) +
1 + strlen(vol->name) + 1) < 0) {
virReportOOMError(conn);
return -1;
}
vol->type = VIR_STORAGE_VOL_FILE;
strcpy(vol->target.path, pool->def->target.path);
strcat(vol->target.path, "/");
strcat(vol->target.path, vol->name);
vol->key = strdup(vol->target.path);
if (vol->key == NULL) {
virReportOOMError(conn);
return -1;
}
if (vol->target.format == VIR_STORAGE_VOL_FILE_RAW) {
if ((fd = open(vol->target.path, O_RDWR | O_CREAT | O_EXCL,
vol->target.perms.mode)) < 0) {
virReportSystemError(conn, errno,
_("cannot create path '%s'"),
vol->target.path);
return -1;
}
/* Pre-allocate any data if requested */
/* XXX slooooooooooooooooow.
* Need to add in progress bars & bg thread somehow */
if (vol->allocation) {
unsigned long long remain = vol->allocation;
static char const zeros[4096];
while (remain) {
int bytes = sizeof(zeros);
if (bytes > remain)
bytes = remain;
if ((bytes = safewrite(fd, zeros, bytes)) < 0) {
virReportSystemError(conn, errno,
_("cannot fill file '%s'"),
vol->target.path);
unlink(vol->target.path);
close(fd);
return -1;
}
remain -= bytes;
}
}
/* Now seek to final size, possibly making the file sparse */
if (ftruncate(fd, vol->capacity) < 0) {
virReportSystemError(conn, errno,
_("cannot extend file '%s'"),
vol->target.path);
unlink(vol->target.path);
close(fd);
return -1;
}
} else if (vol->target.format == VIR_STORAGE_VOL_FILE_DIR) {
if (mkdir(vol->target.path, vol->target.perms.mode) < 0) {
virReportSystemError(conn, errno,
_("cannot create path '%s'"),
vol->target.path);
return -1;
}
if ((fd = open(vol->target.path, O_RDWR)) < 0) {
virReportSystemError(conn, errno,
_("cannot read path '%s'"),
vol->target.path);
return -1;
}
} else {
#if HAVE_QEMU_IMG
const char *type = virStorageVolFormatFileSystemTypeToString(vol->target.format);
const char *backingType = vol->backingStore.path ?
virStorageVolFormatFileSystemTypeToString(vol->backingStore.format) : NULL;
char size[100];
const char **imgargv;
const char *imgargvnormal[] = {
QEMU_IMG, "create", "-f", type, vol->target.path, size, NULL,
};
/* XXX including "backingType" here too, once QEMU accepts
* the patches to specify it. It'll probably be -F backingType */
const char *imgargvbacking[] = {
QEMU_IMG, "create", "-f", type, "-b", vol->backingStore.path, vol->target.path, size, NULL,
};
if (type == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
_("unknown storage vol type %d"),
vol->target.format);
return -1;
}
if (vol->backingStore.path == NULL) {
imgargv = imgargvnormal;
} else {
if (backingType == NULL) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
_("unknown storage vol backing store type %d"),
vol->backingStore.format);
return -1;
}
if (access(vol->backingStore.path, R_OK) != 0) {
virReportSystemError(conn, errno,
_("inaccessible backing store volume %s"),
vol->backingStore.path);
return -1;
}
imgargv = imgargvbacking;
}
/* Size in KB */
snprintf(size, sizeof(size), "%llu", vol->capacity/1024);
if (virRun(conn, imgargv, NULL) < 0) {
unlink(vol->target.path);
return -1;
}
if ((fd = open(vol->target.path, O_RDONLY)) < 0) {
virReportSystemError(conn, errno,
_("cannot read path '%s'"),
vol->target.path);
unlink(vol->target.path);
return -1;
}
#elif HAVE_QCOW_CREATE
/*
* Xen removed the fully-functional qemu-img, and replaced it
* with a partially functional qcow-create. Go figure ??!?
*/
char size[100];
const char *imgargv[4];
if (vol->target.format != VIR_STORAGE_VOL_FILE_QCOW2) {
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
_("unsupported storage vol type %d"),
vol->target.format);
return -1;
}
if (vol->backingStore.path != NULL) {
virStorageReportError(conn, VIR_ERR_NO_SUPPORT,
_("copy-on-write image not supported with "
"qcow-create"));
return -1;
}
/* Size in MB - yes different units to qemu-img :-( */
snprintf(size, sizeof(size), "%llu", vol->capacity/1024/1024);
imgargv[0] = QCOW_CREATE;
imgargv[1] = size;
imgargv[2] = vol->target.path;
imgargv[3] = NULL;
if (virRun(conn, imgargv, NULL) < 0) {
unlink(vol->target.path);
return -1;
}
if ((fd = open(vol->target.path, O_RDONLY)) < 0) {
virReportSystemError(conn, errno,
_("cannot read path '%s'"),
vol->target.path);
unlink(vol->target.path);
return -1;
}
#else
virStorageReportError(conn, VIR_ERR_INTERNAL_ERROR,
"%s", _("creation of non-raw images "
"is not supported without qemu-img"));
return -1;
#endif
}
/* We can only chown/grp if root */
if (getuid() == 0) {
if (fchown(fd, vol->target.perms.uid, vol->target.perms.gid) < 0) {
virReportSystemError(conn, errno,
_("cannot set file owner '%s'"),
vol->target.path);
unlink(vol->target.path);
close(fd);
return -1;
}
}
if (fchmod(fd, vol->target.perms.mode) < 0) {
virReportSystemError(conn, errno,
_("cannot set file mode '%s'"),
vol->target.path);
unlink(vol->target.path);
close(fd);
return -1;
}
/* Refresh allocation / permissions info, but not capacity */
if (virStorageBackendUpdateVolTargetInfoFD(conn, &vol->target, fd,
&vol->allocation,
NULL) < 0) {
unlink(vol->target.path);
close(fd);
return -1;
}
if (close(fd) < 0) {
virReportSystemError(conn, errno,
_("cannot close file '%s'"),
vol->target.path);
unlink(vol->target.path);
return -1;
}
return 0;
}
/**
* Remove a volume - just unlinks for now
*/
static int
virStorageBackendFileSystemVolDelete(virConnectPtr conn,
virStoragePoolObjPtr pool ATTRIBUTE_UNUSED,
virStorageVolDefPtr vol,
unsigned int flags ATTRIBUTE_UNUSED)
{
if (unlink(vol->target.path) < 0) {
/* Silently ignore failures where the vol has already gone away */
if (errno != ENOENT) {
virReportSystemError(conn, errno,
_("cannot unlink file '%s'"),
vol->target.path);
return -1;
}
}
return 0;
}
/**
* Update info about a volume's capacity/allocation
*/
static int
virStorageBackendFileSystemVolRefresh(virConnectPtr conn,
virStoragePoolObjPtr pool ATTRIBUTE_UNUSED,
virStorageVolDefPtr vol)
{
/* Refresh allocation / permissions info in case its changed */
return virStorageBackendUpdateVolInfo(conn, vol, 0);
}
virStorageBackend virStorageBackendDirectory = {
.type = VIR_STORAGE_POOL_DIR,
.buildPool = virStorageBackendFileSystemBuild,
.refreshPool = virStorageBackendFileSystemRefresh,
.deletePool = virStorageBackendFileSystemDelete,
.createVol = virStorageBackendFileSystemVolCreate,
.refreshVol = virStorageBackendFileSystemVolRefresh,
.deleteVol = virStorageBackendFileSystemVolDelete,
};
#if WITH_STORAGE_FS
virStorageBackend virStorageBackendFileSystem = {
.type = VIR_STORAGE_POOL_FS,
.buildPool = virStorageBackendFileSystemBuild,
.startPool = virStorageBackendFileSystemStart,
.refreshPool = virStorageBackendFileSystemRefresh,
.stopPool = virStorageBackendFileSystemStop,
.deletePool = virStorageBackendFileSystemDelete,
.createVol = virStorageBackendFileSystemVolCreate,
.refreshVol = virStorageBackendFileSystemVolRefresh,
.deleteVol = virStorageBackendFileSystemVolDelete,
};
virStorageBackend virStorageBackendNetFileSystem = {
.type = VIR_STORAGE_POOL_NETFS,
.buildPool = virStorageBackendFileSystemBuild,
.startPool = virStorageBackendFileSystemStart,
.findPoolSources = virStorageBackendFileSystemNetFindPoolSources,
.refreshPool = virStorageBackendFileSystemRefresh,
.stopPool = virStorageBackendFileSystemStop,
.deletePool = virStorageBackendFileSystemDelete,
.createVol = virStorageBackendFileSystemVolCreate,
.refreshVol = virStorageBackendFileSystemVolRefresh,
.deleteVol = virStorageBackendFileSystemVolDelete,
};
#endif /* WITH_STORAGE_FS */