/* * storage_backend_fs.c: storage backend for FS and directory handling * * Copyright (C) 2007-2015 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, see * . * * Author: Daniel P. Berrange */ #include #include #include #include #include #include #include #include #include #include #include #include #include #if WITH_BLKID # include #endif #include "virerror.h" #include "storage_backend_fs.h" #include "storage_conf.h" #include "virstoragefile.h" #include "vircommand.h" #include "viralloc.h" #include "virxml.h" #include "virfile.h" #include "virlog.h" #include "virstring.h" #define VIR_FROM_THIS VIR_FROM_STORAGE VIR_LOG_INIT("storage.storage_backend_fs"); #define VIR_STORAGE_VOL_FS_OPEN_FLAGS (VIR_STORAGE_VOL_OPEN_DEFAULT | \ VIR_STORAGE_VOL_OPEN_DIR) #define VIR_STORAGE_VOL_FS_PROBE_FLAGS (VIR_STORAGE_VOL_FS_OPEN_FLAGS | \ VIR_STORAGE_VOL_OPEN_NOERROR) static int virStorageBackendProbeTarget(virStorageSourcePtr target, virStorageEncryptionPtr *encryption) { int backingStoreFormat; int fd = -1; int ret = -1; int rc; virStorageSourcePtr meta = NULL; struct stat sb; if (encryption) *encryption = NULL; if ((rc = virStorageBackendVolOpen(target->path, &sb, VIR_STORAGE_VOL_FS_PROBE_FLAGS)) < 0) return rc; /* Take care to propagate rc, it is not always -1 */ fd = rc; if (virStorageBackendUpdateVolTargetInfoFD(target, fd, &sb) < 0) goto cleanup; if (S_ISDIR(sb.st_mode)) { target->format = VIR_STORAGE_FILE_DIR; ret = 0; goto cleanup; } if (!(meta = virStorageFileGetMetadataFromFD(target->path, fd, VIR_STORAGE_FILE_AUTO, &backingStoreFormat))) goto cleanup; if (meta->backingStoreRaw) { if (!(target->backingStore = virStorageSourceNewFromBacking(meta))) goto cleanup; target->backingStore->format = backingStoreFormat; /* XXX: Remote storage doesn't play nicely with volumes backed by * remote storage. To avoid trouble, just fake the backing store is RAW * and put the string from the metadata as the path of the target. */ if (!virStorageSourceIsLocalStorage(target->backingStore)) { virStorageSourceFree(target->backingStore); if (VIR_ALLOC(target->backingStore) < 0) goto cleanup; target->backingStore->type = VIR_STORAGE_TYPE_NETWORK; target->backingStore->path = meta->backingStoreRaw; meta->backingStoreRaw = NULL; target->backingStore->format = VIR_STORAGE_FILE_RAW; } if (target->backingStore->format == VIR_STORAGE_FILE_AUTO) { if ((rc = virStorageFileProbeFormat(target->backingStore->path, -1, -1)) < 0) { /* If the backing file is currently unavailable or is * accessed via remote protocol only log an error, fake the * format as RAW and continue. Returning -1 here would * disable the whole storage pool, making it unavailable for * even maintenance. */ target->backingStore->format = VIR_STORAGE_FILE_RAW; virReportError(VIR_ERR_INTERNAL_ERROR, _("cannot probe backing volume format: %s"), target->backingStore->path); } else { target->backingStore->format = rc; } } } target->format = meta->format; /* Default to success below this point */ ret = 0; if (meta->capacity) target->capacity = meta->capacity; if (encryption && meta->encryption) { *encryption = meta->encryption; meta->encryption = NULL; switch (target->format) { case VIR_STORAGE_FILE_QCOW: case VIR_STORAGE_FILE_QCOW2: (*encryption)->format = VIR_STORAGE_ENCRYPTION_FORMAT_QCOW; break; default: break; } /* XXX ideally we'd fill in secret UUID here * but we cannot guarantee 'conn' is non-NULL * at this point in time :-( So we only fill * in secrets when someone first queries a vol */ } virBitmapFree(target->features); target->features = meta->features; meta->features = NULL; if (meta->compat) { VIR_FREE(target->compat); target->compat = meta->compat; meta->compat = NULL; } cleanup: VIR_FORCE_CLOSE(fd); virStorageSourceFree(meta); return ret; } #if WITH_STORAGE_FS # include struct _virNetfsDiscoverState { const char *host; virStoragePoolSourceList list; }; typedef struct _virNetfsDiscoverState virNetfsDiscoverState; static int virStorageBackendFileSystemNetFindPoolSourcesFunc(char **const groups, void *data) { virNetfsDiscoverState *state = data; const char *name, *path; virStoragePoolSource *src = NULL; int ret = -1; path = groups[0]; if (!(name = strrchr(path, '/'))) { virReportError(VIR_ERR_INTERNAL_ERROR, _("invalid netfs path (no /): %s"), path); goto cleanup; } name += 1; if (*name == '\0') { virReportError(VIR_ERR_INTERNAL_ERROR, _("invalid netfs path (ends in /): %s"), path); goto cleanup; } if (!(src = virStoragePoolSourceListNewSource(&state->list))) goto cleanup; if (VIR_ALLOC_N(src->hosts, 1) < 0) goto cleanup; src->nhost = 1; if (VIR_STRDUP(src->hosts[0].name, state->host) < 0 || VIR_STRDUP(src->dir, path) < 0) goto cleanup; src->format = VIR_STORAGE_POOL_NETFS_NFS; ret = 0; cleanup: return ret; } static int virStorageBackendFileSystemNetFindNFSPoolSources(virNetfsDiscoverState *state) { int ret = -1; /* * # 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 }; virCommandPtr cmd = NULL; cmd = virCommandNewArgList(SHOWMOUNT, "--no-headers", "--exports", state->host, NULL); if (virCommandRunRegex(cmd, 1, regexes, vars, virStorageBackendFileSystemNetFindPoolSourcesFunc, state, NULL) < 0) goto cleanup; ret = 0; cleanup: virCommandFree(cmd); return ret; } static char * virStorageBackendFileSystemNetFindPoolSources(virConnectPtr conn ATTRIBUTE_UNUSED, const char *srcSpec, unsigned int flags) { virNetfsDiscoverState state = { .host = NULL, .list = { .type = VIR_STORAGE_POOL_NETFS, .nsources = 0, .sources = NULL } }; virStoragePoolSourcePtr source = NULL; char *ret = NULL; size_t i; int retNFS = -1, retGluster = -1; virCheckFlags(0, NULL); if (!srcSpec) { virReportError(VIR_ERR_INVALID_ARG, "%s", _("hostname must be specified for netfs sources")); return NULL; } if (!(source = virStoragePoolDefParseSourceString(srcSpec, VIR_STORAGE_POOL_NETFS))) return NULL; if (source->nhost != 1) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("Expected exactly 1 host for the storage pool")); goto cleanup; } state.host = source->hosts[0].name; retNFS = virStorageBackendFileSystemNetFindNFSPoolSources(&state); # ifdef GLUSTER_CLI retGluster = virStorageBackendFindGlusterPoolSources(state.host, VIR_STORAGE_POOL_NETFS_GLUSTERFS, &state.list); # endif /* If both fail, then we won't return an empty list - return an error */ if (retNFS < 0 && retGluster < 0) goto cleanup; if (!(ret = virStoragePoolSourceListFormat(&state.list))) goto cleanup; cleanup: for (i = 0; i < state.list.nsources; i++) virStoragePoolSourceClear(&state.list.sources[i]); VIR_FREE(state.list.sources); virStoragePoolSourceFree(source); return ret; } /** * @pool storage pool to check FS types * * Determine if storage pool FS types are properly set up * * Return 0 if everything's OK, -1 on error */ static int virStorageBackendFileSystemIsValid(virStoragePoolObjPtr pool) { if (pool->def->type == VIR_STORAGE_POOL_NETFS) { if (pool->def->source.nhost != 1) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("expected exactly 1 host for the storage pool")); return -1; } if (pool->def->source.hosts[0].name == NULL) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing source host")); return -1; } if (pool->def->source.dir == NULL) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing source path")); return -1; } } else { if (pool->def->source.ndevice != 1) { if (pool->def->source.ndevice == 0) virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("missing source device")); else virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("expected exactly 1 device for the " "storage pool")); return -1; } } return 0; } /** * virStorageBackendFileSystemGetPoolSource * @pool: storage pool object pointer * * Allocate/return a string representing the FS storage pool source. * It is up to the caller to VIR_FREE the allocated string */ static char * virStorageBackendFileSystemGetPoolSource(virStoragePoolObjPtr pool) { char *src = NULL; if (pool->def->type == VIR_STORAGE_POOL_NETFS) { if (pool->def->source.format == VIR_STORAGE_POOL_NETFS_CIFS) { if (virAsprintf(&src, "//%s/%s", pool->def->source.hosts[0].name, pool->def->source.dir) < 0) return NULL; } else { if (virAsprintf(&src, "%s:%s", pool->def->source.hosts[0].name, pool->def->source.dir) < 0) return NULL; } } else { if (VIR_STRDUP(src, pool->def->source.devices[0].path) < 0) return NULL; } return src; } /** * @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(virStoragePoolObjPtr pool) { int ret = -1; char *src = NULL; FILE *mtab; struct mntent ent; char buf[1024]; if ((mtab = fopen(_PATH_MOUNTED, "r")) == NULL) { virReportSystemError(errno, _("cannot read mount list '%s'"), _PATH_MOUNTED); goto cleanup; } while ((getmntent_r(mtab, &ent, buf, sizeof(buf))) != NULL) { if (!(src = virStorageBackendFileSystemGetPoolSource(pool))) goto cleanup; if (STREQ(ent.mnt_dir, pool->def->target.path) && STREQ(ent.mnt_fsname, src)) { ret = 1; goto cleanup; } VIR_FREE(src); } ret = 0; cleanup: VIR_FORCE_FCLOSE(mtab); VIR_FREE(src); return ret; } /** * @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(virStoragePoolObjPtr pool) { char *src = NULL; /* '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 */ bool netauto = (pool->def->type == VIR_STORAGE_POOL_NETFS && pool->def->source.format == VIR_STORAGE_POOL_NETFS_AUTO); bool glusterfs = (pool->def->type == VIR_STORAGE_POOL_NETFS && pool->def->source.format == VIR_STORAGE_POOL_NETFS_GLUSTERFS); bool cifsfs = (pool->def->type == VIR_STORAGE_POOL_NETFS && pool->def->source.format == VIR_STORAGE_POOL_NETFS_CIFS); virCommandPtr cmd = NULL; int ret = -1; int rc; if (virStorageBackendFileSystemIsValid(pool) < 0) return -1; /* Short-circuit if already mounted */ if ((rc = virStorageBackendFileSystemIsMounted(pool)) != 0) { if (rc == 1) { virReportError(VIR_ERR_OPERATION_INVALID, _("Target '%s' is already mounted"), pool->def->target.path); } return -1; } if (!(src = virStorageBackendFileSystemGetPoolSource(pool))) return -1; if (netauto) cmd = virCommandNewArgList(MOUNT, src, pool->def->target.path, NULL); else if (glusterfs) cmd = virCommandNewArgList(MOUNT, "-t", virStoragePoolFormatFileSystemNetTypeToString(pool->def->source.format), src, "-o", "direct-io-mode=1", pool->def->target.path, NULL); else if (cifsfs) cmd = virCommandNewArgList(MOUNT, "-t", virStoragePoolFormatFileSystemNetTypeToString(pool->def->source.format), src, pool->def->target.path, "-o", "guest", NULL); else cmd = virCommandNewArgList(MOUNT, "-t", (pool->def->type == VIR_STORAGE_POOL_FS ? virStoragePoolFormatFileSystemTypeToString(pool->def->source.format) : virStoragePoolFormatFileSystemNetTypeToString(pool->def->source.format)), src, pool->def->target.path, NULL); if (virCommandRun(cmd, NULL) < 0) goto cleanup; ret = 0; cleanup: virCommandFree(cmd); VIR_FREE(src); return ret; } /** * @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(virStoragePoolObjPtr pool) { virCommandPtr cmd = NULL; int ret = -1; int rc; if (virStorageBackendFileSystemIsValid(pool) < 0) return -1; /* Short-circuit if already unmounted */ if ((rc = virStorageBackendFileSystemIsMounted(pool)) != 1) return rc; cmd = virCommandNewArgList(UMOUNT, pool->def->target.path, NULL); if (virCommandRun(cmd, NULL) < 0) goto cleanup; ret = 0; cleanup: virCommandFree(cmd); return ret; } #endif /* WITH_STORAGE_FS */ static int virStorageBackendFileSystemCheck(virStoragePoolObjPtr pool, bool *isActive) { if (pool->def->type == VIR_STORAGE_POOL_DIR) { *isActive = virFileExists(pool->def->target.path); #if WITH_STORAGE_FS } else { int ret; *isActive = false; if (virStorageBackendFileSystemIsValid(pool) < 0) return -1; if ((ret = virStorageBackendFileSystemIsMounted(pool)) != 0) { if (ret < 0) return -1; *isActive = true; } #endif /* WITH_STORAGE_FS */ } return 0; } #if WITH_STORAGE_FS /** * @conn connection to report errors against * @pool storage pool to start * * Starts a directory or FS based storage pool. The underlying source * device will be mounted for FS based pools. * * Returns 0 on success, -1 on error */ static int virStorageBackendFileSystemStart(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool) { if (pool->def->type != VIR_STORAGE_POOL_DIR && virStorageBackendFileSystemMount(pool) < 0) return -1; return 0; } #endif /* WITH_STORAGE_FS */ #if WITH_BLKID static virStoragePoolProbeResult virStorageBackendFileSystemProbe(const char *device, const char *format) { virStoragePoolProbeResult ret = FILESYSTEM_PROBE_ERROR; blkid_probe probe = NULL; const char *fstype = NULL; char *names[2], *libblkid_format = NULL; VIR_DEBUG("Probing for existing filesystem of type %s on device %s", format, device); if (blkid_known_fstype(format) == 0) { virReportError(VIR_ERR_STORAGE_PROBE_FAILED, _("Not capable of probing for " "filesystem of type %s"), format); goto error; } probe = blkid_new_probe_from_filename(device); if (probe == NULL) { virReportError(VIR_ERR_STORAGE_PROBE_FAILED, _("Failed to create filesystem probe " "for device %s"), device); goto error; } if (VIR_STRDUP(libblkid_format, format) < 0) goto error; names[0] = libblkid_format; names[1] = NULL; blkid_probe_filter_superblocks_type(probe, BLKID_FLTR_ONLYIN, names); if (blkid_do_probe(probe) != 0) { VIR_INFO("No filesystem of type '%s' found on device '%s'", format, device); ret = FILESYSTEM_PROBE_NOT_FOUND; } else if (blkid_probe_lookup_value(probe, "TYPE", &fstype, NULL) == 0) { virReportError(VIR_ERR_STORAGE_POOL_BUILT, _("Existing filesystem of type '%s' found on " "device '%s'"), fstype, device); ret = FILESYSTEM_PROBE_FOUND; } if (blkid_do_probe(probe) != 1) { virReportError(VIR_ERR_STORAGE_PROBE_FAILED, "%s", _("Found additional probes to run, " "filesystem probing may be incorrect")); ret = FILESYSTEM_PROBE_ERROR; } error: VIR_FREE(libblkid_format); if (probe != NULL) blkid_free_probe(probe); return ret; } #else /* #if WITH_BLKID */ static virStoragePoolProbeResult virStorageBackendFileSystemProbe(const char *device ATTRIBUTE_UNUSED, const char *format ATTRIBUTE_UNUSED) { virReportError(VIR_ERR_OPERATION_INVALID, "%s", _("probing for filesystems is unsupported " "by this build")); return FILESYSTEM_PROBE_ERROR; } #endif /* #if WITH_BLKID */ /* some platforms don't support mkfs */ #ifdef MKFS static int virStorageBackendExecuteMKFS(const char *device, const char *format) { int ret = 0; virCommandPtr cmd = NULL; cmd = virCommandNewArgList(MKFS, "-t", format, NULL); /* use the force, otherwise mkfs.xfs won't overwrite existing fs */ if (STREQ(format, "xfs")) virCommandAddArg(cmd, "-f"); virCommandAddArg(cmd, device); if (virCommandRun(cmd, NULL) < 0) { virReportSystemError(errno, _("Failed to make filesystem of " "type '%s' on device '%s'"), format, device); ret = -1; } virCommandFree(cmd); return ret; } #else /* #ifdef MKFS */ static int virStorageBackendExecuteMKFS(const char *device ATTRIBUTE_UNUSED, const char *format ATTRIBUTE_UNUSED) { virReportError(VIR_ERR_INTERNAL_ERROR, _("mkfs is not supported on this platform: " "Failed to make filesystem of " "type '%s' on device '%s'"), format, device); return -1; } #endif /* #ifdef MKFS */ static int virStorageBackendMakeFileSystem(virStoragePoolObjPtr pool, unsigned int flags) { const char *device = NULL, *format = NULL; bool ok_to_mkfs = false; int ret = -1; if (pool->def->source.devices == NULL) { virReportError(VIR_ERR_OPERATION_INVALID, _("No source device specified when formatting pool '%s'"), pool->def->name); goto error; } device = pool->def->source.devices[0].path; format = virStoragePoolFormatFileSystemTypeToString(pool->def->source.format); VIR_DEBUG("source device: '%s' format: '%s'", device, format); if (!virFileExists(device)) { virReportError(VIR_ERR_OPERATION_INVALID, _("Source device does not exist when formatting pool '%s'"), pool->def->name); goto error; } if (flags & VIR_STORAGE_POOL_BUILD_OVERWRITE) { ok_to_mkfs = true; } else if (flags & VIR_STORAGE_POOL_BUILD_NO_OVERWRITE && virStorageBackendFileSystemProbe(device, format) == FILESYSTEM_PROBE_NOT_FOUND) { ok_to_mkfs = true; } if (ok_to_mkfs) ret = virStorageBackendExecuteMKFS(device, format); error: return ret; } /** * @conn connection to report errors against * @pool storage pool to build * @flags controls the pool formatting behaviour * * Build a directory or FS based storage pool. * * If no flag is set, it only makes the directory; If * VIR_STORAGE_POOL_BUILD_NO_OVERWRITE set, it probes to determine if * filesystem already exists on the target device, renurning an error * if exists, or using mkfs to format the target device if not; If * VIR_STORAGE_POOL_BUILD_OVERWRITE is set, mkfs is always executed, * any existed data on the target device is overwritten unconditionally. * * The underlying source device is mounted for FS based pools. * * Returns 0 on success, -1 on error */ static int virStorageBackendFileSystemBuild(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool, unsigned int flags) { int ret = -1; char *parent = NULL; char *p = NULL; mode_t mode; bool needs_create_as_uid; unsigned int dir_create_flags; virCheckFlags(VIR_STORAGE_POOL_BUILD_OVERWRITE | VIR_STORAGE_POOL_BUILD_NO_OVERWRITE, ret); VIR_EXCLUSIVE_FLAGS_GOTO(VIR_STORAGE_POOL_BUILD_OVERWRITE, VIR_STORAGE_POOL_BUILD_NO_OVERWRITE, error); if (VIR_STRDUP(parent, pool->def->target.path) < 0) goto error; if (!(p = strrchr(parent, '/'))) { virReportError(VIR_ERR_INVALID_ARG, _("path '%s' is not absolute"), pool->def->target.path); goto error; } if (p != parent) { /* assure all directories in the path prior to the final dir * exist, with default uid/gid/mode. */ *p = '\0'; if (virFileMakePath(parent) < 0) { virReportSystemError(errno, _("cannot create path '%s'"), parent); goto error; } } dir_create_flags = VIR_DIR_CREATE_ALLOW_EXIST; needs_create_as_uid = (pool->def->type == VIR_STORAGE_POOL_NETFS); mode = pool->def->target.perms.mode; if (mode == (mode_t) -1 && (needs_create_as_uid || !virFileExists(pool->def->target.path))) mode = VIR_STORAGE_DEFAULT_POOL_PERM_MODE; if (needs_create_as_uid) dir_create_flags |= VIR_DIR_CREATE_AS_UID; /* Now create the final dir in the path with the uid/gid/mode * requested in the config. If the dir already exists, just set * the perms. */ if (virDirCreate(pool->def->target.path, mode, pool->def->target.perms.uid, pool->def->target.perms.gid, dir_create_flags) < 0) goto error; if (flags != 0) { ret = virStorageBackendMakeFileSystem(pool, flags); } else { ret = 0; } error: VIR_FREE(parent); return ret; } /** * Iterate over the pool's directory and enumerate all disk images * within it. This is non-recursive. */ static int virStorageBackendFileSystemRefresh(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool) { DIR *dir; struct dirent *ent; struct statvfs sb; struct stat statbuf; virStorageVolDefPtr vol = NULL; virStorageSourcePtr target = NULL; int direrr; int fd = -1, ret = -1; if (!(dir = opendir(pool->def->target.path))) { virReportSystemError(errno, _("cannot open path '%s'"), pool->def->target.path); goto cleanup; } while ((direrr = virDirRead(dir, &ent, pool->def->target.path)) > 0) { int err; if (virStringHasControlChars(ent->d_name)) { VIR_WARN("Ignoring file with control characters under '%s'", pool->def->target.path); continue; } if (VIR_ALLOC(vol) < 0) goto cleanup; if (VIR_STRDUP(vol->name, ent->d_name) < 0) goto cleanup; vol->type = VIR_STORAGE_VOL_FILE; vol->target.format = VIR_STORAGE_FILE_RAW; /* Real value is filled in during probe */ if (virAsprintf(&vol->target.path, "%s/%s", pool->def->target.path, vol->name) == -1) goto cleanup; if (VIR_STRDUP(vol->key, vol->target.path) < 0) goto cleanup; if ((err = virStorageBackendProbeTarget(&vol->target, &vol->target.encryption)) < 0) { if (err == -2) { /* Silently ignore non-regular files, * eg '.' '..', 'lost+found', dangling symbolic link */ virStorageVolDefFree(vol); vol = NULL; continue; } else if (err == -3) { /* The backing file is currently unavailable, its format is not * explicitly specified, the probe to auto detect the format * failed: continue with faked RAW format, since AUTO will * break virStorageVolTargetDefFormat() generating the line * . */ } else { goto cleanup; } } /* directory based volume */ if (vol->target.format == VIR_STORAGE_FILE_DIR) vol->type = VIR_STORAGE_VOL_DIR; if (vol->target.backingStore) { ignore_value(virStorageBackendUpdateVolTargetInfo(vol->target.backingStore, false, VIR_STORAGE_VOL_OPEN_DEFAULT, 0)); /* If this failed, the backing file is currently unavailable, * the capacity, allocation, owner, group and mode are unknown. * An error message was raised, but we just continue. */ } if (VIR_APPEND_ELEMENT(pool->volumes.objs, pool->volumes.count, vol) < 0) goto cleanup; } if (direrr < 0) goto cleanup; closedir(dir); dir = NULL; vol = NULL; if (VIR_ALLOC(target)) goto cleanup; if ((fd = open(pool->def->target.path, O_RDONLY)) < 0) { virReportSystemError(errno, _("cannot open path '%s'"), pool->def->target.path); goto cleanup; } if (fstat(fd, &statbuf) < 0) { virReportSystemError(errno, _("cannot stat path '%s'"), pool->def->target.path); goto cleanup; } if (virStorageBackendUpdateVolTargetInfoFD(target, fd, &statbuf) < 0) goto cleanup; /* VolTargetInfoFD doesn't update capacity correctly for the pool case */ if (statvfs(pool->def->target.path, &sb) < 0) { virReportSystemError(errno, _("cannot statvfs path '%s'"), pool->def->target.path); goto cleanup; } 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_frsize); pool->def->allocation = pool->def->capacity - pool->def->available; pool->def->target.perms.mode = target->perms->mode; pool->def->target.perms.uid = target->perms->uid; pool->def->target.perms.gid = target->perms->gid; VIR_FREE(pool->def->target.perms.label); if (VIR_STRDUP(pool->def->target.perms.label, target->perms->label) < 0) goto cleanup; ret = 0; cleanup: if (dir) closedir(dir); VIR_FORCE_CLOSE(fd); virStorageVolDefFree(vol); virStorageSourceFree(target); if (ret < 0) virStoragePoolObjClearVols(pool); return ret; } /** * @conn connection to report errors against * @pool storage pool to stop * * Stops a file storage pool. The underlying source device is unmounted * for FS based pools. Any cached data about volumes is released. * * Returns 0 on success, -1 on error. */ #if WITH_STORAGE_FS static int virStorageBackendFileSystemStop(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool) { if (virStorageBackendFileSystemUnmount(pool) < 0) return -1; return 0; } #endif /* WITH_STORAGE_FS */ /** * @conn connection to report errors against * @pool storage pool to delete * * Delete a directory based storage pool * * Returns 0 on success, -1 on error */ static int virStorageBackendFileSystemDelete(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool, unsigned int flags) { virCheckFlags(0, -1); /* XXX delete all vols first ? */ if (rmdir(pool->def->target.path) < 0) { virReportSystemError(errno, _("failed to remove pool '%s'"), pool->def->target.path); return -1; } return 0; } /** * Set up a volume definition to be added to a pool's volume list, but * don't do any file creation or allocation. By separating the two processes, * we allow allocation progress reporting (by polling the volume's 'info' * function), and can drop the parent pool lock during the (slow) allocation. */ static int virStorageBackendFileSystemVolCreate(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool, virStorageVolDefPtr vol) { if (vol->target.format == VIR_STORAGE_FILE_DIR) vol->type = VIR_STORAGE_VOL_DIR; else if (vol->target.format == VIR_STORAGE_FILE_PLOOP) vol->type = VIR_STORAGE_VOL_PLOOP; else vol->type = VIR_STORAGE_VOL_FILE; /* Volumes within a directory pools are not recursive; do not * allow escape to ../ or a subdir */ if (strchr(vol->name, '/')) { virReportError(VIR_ERR_OPERATION_INVALID, _("volume name '%s' cannot contain '/'"), vol->name); return -1; } VIR_FREE(vol->target.path); if (virAsprintf(&vol->target.path, "%s/%s", pool->def->target.path, vol->name) == -1) return -1; if (virFileExists(vol->target.path)) { virReportError(VIR_ERR_OPERATION_INVALID, _("volume target path '%s' already exists"), vol->target.path); return -1; } VIR_FREE(vol->key); return VIR_STRDUP(vol->key, vol->target.path); } static int createFileDir(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool, virStorageVolDefPtr vol, virStorageVolDefPtr inputvol, unsigned int flags) { int err; virCheckFlags(0, -1); if (inputvol) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("cannot copy from volume to a directory volume")); return -1; } if (vol->target.backingStore) { virReportError(VIR_ERR_NO_SUPPORT, "%s", _("backing storage not supported for directories volumes")); return -1; } if ((err = virDirCreate(vol->target.path, (vol->target.perms->mode == (mode_t) -1 ? VIR_STORAGE_DEFAULT_VOL_PERM_MODE : vol->target.perms->mode), vol->target.perms->uid, vol->target.perms->gid, (pool->def->type == VIR_STORAGE_POOL_NETFS ? VIR_DIR_CREATE_AS_UID : 0))) < 0) { return -1; } return 0; } static int _virStorageBackendFileSystemVolBuild(virConnectPtr conn, virStoragePoolObjPtr pool, virStorageVolDefPtr vol, virStorageVolDefPtr inputvol, unsigned int flags) { virStorageBackendBuildVolFrom create_func; int tool_type; if (inputvol) { if (vol->target.encryption != NULL) { virReportError(VIR_ERR_CONFIG_UNSUPPORTED, "%s", _("storage pool does not support " "building encrypted volumes from " "other volumes")); return -1; } create_func = virStorageBackendGetBuildVolFromFunction(vol, inputvol); if (!create_func) return -1; } else if (vol->target.format == VIR_STORAGE_FILE_RAW) { create_func = virStorageBackendCreateRaw; } else if (vol->target.format == VIR_STORAGE_FILE_DIR) { create_func = createFileDir; } else if (vol->target.format == VIR_STORAGE_FILE_PLOOP) { create_func = virStorageBackendCreatePloop; } else if ((tool_type = virStorageBackendFindFSImageTool(NULL)) != -1) { create_func = virStorageBackendFSImageToolTypeToFunc(tool_type); if (!create_func) return -1; } else { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("creation of non-raw images " "is not supported without qemu-img")); return -1; } if (create_func(conn, pool, vol, inputvol, flags) < 0) 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 virStorageBackendFileSystemVolBuild(virConnectPtr conn, virStoragePoolObjPtr pool, virStorageVolDefPtr vol, unsigned int flags) { virCheckFlags(VIR_STORAGE_VOL_CREATE_PREALLOC_METADATA | VIR_STORAGE_VOL_CREATE_REFLINK, -1); return _virStorageBackendFileSystemVolBuild(conn, pool, vol, NULL, flags); } /* * Create a storage vol using 'inputvol' as input */ static int virStorageBackendFileSystemVolBuildFrom(virConnectPtr conn, virStoragePoolObjPtr pool, virStorageVolDefPtr vol, virStorageVolDefPtr inputvol, unsigned int flags) { virCheckFlags(VIR_STORAGE_VOL_CREATE_PREALLOC_METADATA | VIR_STORAGE_VOL_CREATE_REFLINK, -1); return _virStorageBackendFileSystemVolBuild(conn, pool, vol, inputvol, flags); } /** * Remove a volume - no support for BLOCK and NETWORK yet */ static int virStorageBackendFileSystemVolDelete(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool ATTRIBUTE_UNUSED, virStorageVolDefPtr vol, unsigned int flags) { virCheckFlags(0, -1); switch ((virStorageVolType) vol->type) { case VIR_STORAGE_VOL_FILE: case VIR_STORAGE_VOL_DIR: if (virFileRemove(vol->target.path, vol->target.perms->uid, vol->target.perms->gid) < 0) { /* Silently ignore failures where the vol has already gone away */ if (errno != ENOENT) { if (vol->type == VIR_STORAGE_VOL_FILE) virReportSystemError(errno, _("cannot unlink file '%s'"), vol->target.path); else virReportSystemError(errno, _("cannot remove directory '%s'"), vol->target.path); return -1; } } break; case VIR_STORAGE_VOL_PLOOP: if (virFileDeleteTree(vol->target.path) < 0) return -1; break; case VIR_STORAGE_VOL_BLOCK: case VIR_STORAGE_VOL_NETWORK: case VIR_STORAGE_VOL_NETDIR: case VIR_STORAGE_VOL_LAST: virReportError(VIR_ERR_NO_SUPPORT, _("removing block or network volumes is not supported: %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) { int ret; /* Refresh allocation / capacity / permissions info in case its changed */ ret = virStorageBackendUpdateVolInfo(vol, false, VIR_STORAGE_VOL_FS_OPEN_FLAGS, 0); if (ret < 0) return ret; /* Load any secrets if possible */ if (vol->target.encryption && vol->target.encryption->format == VIR_STORAGE_ENCRYPTION_FORMAT_QCOW && vol->target.encryption->nsecrets == 0) { virSecretPtr sec; virStorageEncryptionSecretPtr encsec = NULL; sec = virSecretLookupByUsage(conn, VIR_SECRET_USAGE_TYPE_VOLUME, vol->target.path); if (sec) { if (VIR_ALLOC_N(vol->target.encryption->secrets, 1) < 0 || VIR_ALLOC(encsec) < 0) { VIR_FREE(vol->target.encryption->secrets); virObjectUnref(sec); return -1; } vol->target.encryption->nsecrets = 1; vol->target.encryption->secrets[0] = encsec; encsec->type = VIR_STORAGE_ENCRYPTION_SECRET_TYPE_PASSPHRASE; virSecretGetUUID(sec, encsec->uuid); virObjectUnref(sec); } } return 0; } static int virStorageBackendFilesystemResizeQemuImg(const char *path, unsigned long long capacity) { int ret = -1; char *img_tool; virCommandPtr cmd = NULL; /* KVM is usually ahead of qemu on features, so try that first */ img_tool = virFindFileInPath("kvm-img"); if (!img_tool) img_tool = virFindFileInPath("qemu-img"); if (!img_tool) { virReportError(VIR_ERR_INTERNAL_ERROR, "%s", _("unable to find kvm-img or qemu-img")); return -1; } /* Round capacity as qemu-img resize errors out on sizes which are not * a multiple of 512 */ capacity = VIR_ROUND_UP(capacity, 512); cmd = virCommandNew(img_tool); virCommandAddArgList(cmd, "resize", path, NULL); virCommandAddArgFormat(cmd, "%llu", capacity); ret = virCommandRun(cmd, NULL); VIR_FREE(img_tool); virCommandFree(cmd); return ret; } /** * Resize a volume */ static int virStorageBackendFileSystemVolResize(virConnectPtr conn ATTRIBUTE_UNUSED, virStoragePoolObjPtr pool ATTRIBUTE_UNUSED, virStorageVolDefPtr vol, unsigned long long capacity, unsigned int flags) { virCheckFlags(VIR_STORAGE_VOL_RESIZE_ALLOCATE | VIR_STORAGE_VOL_RESIZE_SHRINK, -1); bool pre_allocate = flags & VIR_STORAGE_VOL_RESIZE_ALLOCATE; if (vol->target.format == VIR_STORAGE_FILE_RAW) { return virStorageFileResize(vol->target.path, capacity, vol->target.allocation, pre_allocate); } else if (vol->target.format == VIR_STORAGE_FILE_PLOOP) { return virStoragePloopResize(vol, capacity); } else { if (pre_allocate) { virReportError(VIR_ERR_OPERATION_UNSUPPORTED, "%s", _("preallocate is only supported for raw " "type volume")); return -1; } return virStorageBackendFilesystemResizeQemuImg(vol->target.path, capacity); } } virStorageBackend virStorageBackendDirectory = { .type = VIR_STORAGE_POOL_DIR, .buildPool = virStorageBackendFileSystemBuild, .checkPool = virStorageBackendFileSystemCheck, .refreshPool = virStorageBackendFileSystemRefresh, .deletePool = virStorageBackendFileSystemDelete, .buildVol = virStorageBackendFileSystemVolBuild, .buildVolFrom = virStorageBackendFileSystemVolBuildFrom, .createVol = virStorageBackendFileSystemVolCreate, .refreshVol = virStorageBackendFileSystemVolRefresh, .deleteVol = virStorageBackendFileSystemVolDelete, .resizeVol = virStorageBackendFileSystemVolResize, .uploadVol = virStorageBackendVolUploadLocal, .downloadVol = virStorageBackendVolDownloadLocal, .wipeVol = virStorageBackendVolWipeLocal, }; #if WITH_STORAGE_FS virStorageBackend virStorageBackendFileSystem = { .type = VIR_STORAGE_POOL_FS, .buildPool = virStorageBackendFileSystemBuild, .checkPool = virStorageBackendFileSystemCheck, .startPool = virStorageBackendFileSystemStart, .refreshPool = virStorageBackendFileSystemRefresh, .stopPool = virStorageBackendFileSystemStop, .deletePool = virStorageBackendFileSystemDelete, .buildVol = virStorageBackendFileSystemVolBuild, .buildVolFrom = virStorageBackendFileSystemVolBuildFrom, .createVol = virStorageBackendFileSystemVolCreate, .refreshVol = virStorageBackendFileSystemVolRefresh, .deleteVol = virStorageBackendFileSystemVolDelete, .resizeVol = virStorageBackendFileSystemVolResize, .uploadVol = virStorageBackendVolUploadLocal, .downloadVol = virStorageBackendVolDownloadLocal, .wipeVol = virStorageBackendVolWipeLocal, }; virStorageBackend virStorageBackendNetFileSystem = { .type = VIR_STORAGE_POOL_NETFS, .buildPool = virStorageBackendFileSystemBuild, .checkPool = virStorageBackendFileSystemCheck, .startPool = virStorageBackendFileSystemStart, .findPoolSources = virStorageBackendFileSystemNetFindPoolSources, .refreshPool = virStorageBackendFileSystemRefresh, .stopPool = virStorageBackendFileSystemStop, .deletePool = virStorageBackendFileSystemDelete, .buildVol = virStorageBackendFileSystemVolBuild, .buildVolFrom = virStorageBackendFileSystemVolBuildFrom, .createVol = virStorageBackendFileSystemVolCreate, .refreshVol = virStorageBackendFileSystemVolRefresh, .deleteVol = virStorageBackendFileSystemVolDelete, .resizeVol = virStorageBackendFileSystemVolResize, .uploadVol = virStorageBackendVolUploadLocal, .downloadVol = virStorageBackendVolDownloadLocal, .wipeVol = virStorageBackendVolWipeLocal, }; typedef struct _virStorageFileBackendFsPriv virStorageFileBackendFsPriv; typedef virStorageFileBackendFsPriv *virStorageFileBackendFsPrivPtr; struct _virStorageFileBackendFsPriv { char *canonpath; /* unique file identifier (canonical path) */ }; static void virStorageFileBackendFileDeinit(virStorageSourcePtr src) { VIR_DEBUG("deinitializing FS storage file %p (%s:%s)", src, virStorageTypeToString(virStorageSourceGetActualType(src)), src->path); virStorageFileBackendFsPrivPtr priv = src->drv->priv; VIR_FREE(priv->canonpath); VIR_FREE(priv); } static int virStorageFileBackendFileInit(virStorageSourcePtr src) { virStorageFileBackendFsPrivPtr priv = NULL; VIR_DEBUG("initializing FS storage file %p (%s:%s)[%u:%u]", src, virStorageTypeToString(virStorageSourceGetActualType(src)), src->path, (unsigned int)src->drv->uid, (unsigned int)src->drv->gid); if (VIR_ALLOC(priv) < 0) return -1; src->drv->priv = priv; return 0; } static int virStorageFileBackendFileCreate(virStorageSourcePtr src) { int fd = -1; mode_t mode = S_IRUSR; if (!src->readonly) mode |= S_IWUSR; if ((fd = virFileOpenAs(src->path, O_WRONLY | O_TRUNC | O_CREAT, mode, src->drv->uid, src->drv->gid, 0)) < 0) { errno = -fd; return -1; } VIR_FORCE_CLOSE(fd); return 0; } static int virStorageFileBackendFileUnlink(virStorageSourcePtr src) { return unlink(src->path); } static int virStorageFileBackendFileStat(virStorageSourcePtr src, struct stat *st) { return stat(src->path, st); } static ssize_t virStorageFileBackendFileReadHeader(virStorageSourcePtr src, ssize_t max_len, char **buf) { int fd = -1; ssize_t ret = -1; if ((fd = virFileOpenAs(src->path, O_RDONLY, 0, src->drv->uid, src->drv->gid, 0)) < 0) { virReportSystemError(-fd, _("Failed to open file '%s'"), src->path); return -1; } if ((ret = virFileReadHeaderFD(fd, max_len, buf)) < 0) { virReportSystemError(errno, _("cannot read header '%s'"), src->path); goto cleanup; } cleanup: VIR_FORCE_CLOSE(fd); return ret; } static const char * virStorageFileBackendFileGetUniqueIdentifier(virStorageSourcePtr src) { virStorageFileBackendFsPrivPtr priv = src->drv->priv; if (!priv->canonpath) { if (!(priv->canonpath = canonicalize_file_name(src->path))) { virReportSystemError(errno, _("can't canonicalize path '%s'"), src->path); return NULL; } } return priv->canonpath; } static int virStorageFileBackendFileAccess(virStorageSourcePtr src, int mode) { return virFileAccessibleAs(src->path, mode, src->drv->uid, src->drv->gid); } static int virStorageFileBackendFileChown(virStorageSourcePtr src, uid_t uid, gid_t gid) { return chown(src->path, uid, gid); } virStorageFileBackend virStorageFileBackendFile = { .type = VIR_STORAGE_TYPE_FILE, .backendInit = virStorageFileBackendFileInit, .backendDeinit = virStorageFileBackendFileDeinit, .storageFileCreate = virStorageFileBackendFileCreate, .storageFileUnlink = virStorageFileBackendFileUnlink, .storageFileStat = virStorageFileBackendFileStat, .storageFileReadHeader = virStorageFileBackendFileReadHeader, .storageFileAccess = virStorageFileBackendFileAccess, .storageFileChown = virStorageFileBackendFileChown, .storageFileGetUniqueIdentifier = virStorageFileBackendFileGetUniqueIdentifier, }; virStorageFileBackend virStorageFileBackendBlock = { .type = VIR_STORAGE_TYPE_BLOCK, .backendInit = virStorageFileBackendFileInit, .backendDeinit = virStorageFileBackendFileDeinit, .storageFileStat = virStorageFileBackendFileStat, .storageFileReadHeader = virStorageFileBackendFileReadHeader, .storageFileAccess = virStorageFileBackendFileAccess, .storageFileChown = virStorageFileBackendFileChown, .storageFileGetUniqueIdentifier = virStorageFileBackendFileGetUniqueIdentifier, }; virStorageFileBackend virStorageFileBackendDir = { .type = VIR_STORAGE_TYPE_DIR, .backendInit = virStorageFileBackendFileInit, .backendDeinit = virStorageFileBackendFileDeinit, .storageFileAccess = virStorageFileBackendFileAccess, .storageFileChown = virStorageFileBackendFileChown, .storageFileGetUniqueIdentifier = virStorageFileBackendFileGetUniqueIdentifier, }; #endif /* WITH_STORAGE_FS */