freebsd-dev/lib/libbe/be_info.c
Kyle Evans 162ec56949 libbe(3): Fix error handling with respect to be_exists
Some paths through be_exists will set the error state, others will not
There are multiple reasons that a call can fail, so clean it up a bit: all
paths now return an appropriate error code so the caller can attempt to
distinguish between a BE legitimately not existing and just having the wrong
mountpoint. The caller is expected to bubble the error through to the
internal error handler as needed.

This fixes some unfriendliness with bectl(8)'s activate subcommand, where
it might fail due to a bad mountpoint but the only message output is a
generic "failed to activate" message.

Approved by:	re (gjb)
2018-09-01 02:22:26 +00:00

319 lines
7.5 KiB
C

/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 2017 Kyle J. Kneitinger <kyle@kneit.in>
* Copyright (c) 2018 Kyle Evans <kevans@FreeBSD.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "be.h"
#include "be_impl.h"
static int snapshot_proplist_update(zfs_handle_t *hdl, prop_data_t *data);
/*
* Returns the name of the active boot environment
*/
const char *
be_active_name(libbe_handle_t *lbh)
{
return (strrchr(lbh->rootfs, '/') + sizeof(char));
}
/*
* Returns full path of the active boot environment
*/
const char *
be_active_path(libbe_handle_t *lbh)
{
return (lbh->rootfs);
}
/*
* Returns the name of the next active boot environment
*/
const char *
be_nextboot_name(libbe_handle_t *lbh)
{
return (strrchr(lbh->bootfs, '/') + sizeof(char));
}
/*
* Returns full path of the active boot environment
*/
const char *
be_nextboot_path(libbe_handle_t *lbh)
{
return (lbh->bootfs);
}
/*
* Returns the path of the boot environment root dataset
*/
const char *
be_root_path(libbe_handle_t *lbh)
{
return (lbh->root);
}
/*
* Populates dsnvl with one nvlist per bootenv dataset describing the properties
* of that dataset that we've declared ourselves to care about.
*/
int
be_get_bootenv_props(libbe_handle_t *lbh, nvlist_t *dsnvl)
{
prop_data_t data;
data.lbh = lbh;
data.list = dsnvl;
data.single_object = false;
return (be_proplist_update(&data));
}
int
be_get_dataset_props(libbe_handle_t *lbh, const char *name, nvlist_t *props)
{
zfs_handle_t *snap_hdl;
prop_data_t data;
int ret;
data.lbh = lbh;
data.list = props;
data.single_object = true;
if ((snap_hdl = zfs_open(lbh->lzh, name,
ZFS_TYPE_FILESYSTEM | ZFS_TYPE_SNAPSHOT)) == NULL)
return (BE_ERR_ZFSOPEN);
ret = prop_list_builder_cb(snap_hdl, &data);
zfs_close(snap_hdl);
return (ret);
}
int
be_get_dataset_snapshots(libbe_handle_t *lbh, const char *name, nvlist_t *props)
{
zfs_handle_t *ds_hdl;
prop_data_t data;
int ret;
data.lbh = lbh;
data.list = props;
data.single_object = false;
if ((ds_hdl = zfs_open(lbh->lzh, name,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (BE_ERR_ZFSOPEN);
ret = snapshot_proplist_update(ds_hdl, &data);
zfs_close(ds_hdl);
return (ret);
}
/*
* Internal callback function used by zfs_iter_filesystems. For each dataset in
* the bootenv root, populate an nvlist_t of its relevant properties.
*/
int
prop_list_builder_cb(zfs_handle_t *zfs_hdl, void *data_p)
{
char buf[512], *mountpoint;
prop_data_t *data;
libbe_handle_t *lbh;
nvlist_t *props;
const char *dataset, *name;
boolean_t mounted;
/*
* XXX TODO:
* some system for defining constants for the nvlist keys
* error checking
*/
data = (prop_data_t *)data_p;
lbh = data->lbh;
if (data->single_object)
props = data->list;
else
nvlist_alloc(&props, NV_UNIQUE_NAME, KM_SLEEP);
dataset = zfs_get_name(zfs_hdl);
nvlist_add_string(props, "dataset", dataset);
name = strrchr(dataset, '/') + 1;
nvlist_add_string(props, "name", name);
mounted = zfs_is_mounted(zfs_hdl, &mountpoint);
if (mounted)
nvlist_add_string(props, "mounted", mountpoint);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_MOUNTPOINT, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "mountpoint", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_ORIGIN, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "origin", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_CREATION, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "creation", buf);
nvlist_add_boolean_value(props, "active",
(strcmp(be_active_path(lbh), dataset) == 0));
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USED, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "used", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDDS, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedds", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDSNAP, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedsnap", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_USEDREFRESERV, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "usedrefreserv", buf);
if (zfs_prop_get(zfs_hdl, ZFS_PROP_REFERENCED, buf, 512,
NULL, NULL, 0, 1) == 0)
nvlist_add_string(props, "referenced", buf);
nvlist_add_boolean_value(props, "nextboot",
(strcmp(be_nextboot_path(lbh), dataset) == 0));
if (!data->single_object)
nvlist_add_nvlist(data->list, name, props);
return (0);
}
/*
* Updates the properties of each bootenv in the libbe handle
* XXX TODO: ensure that this is always consistent (run after adds, deletes,
* renames,etc
*/
int
be_proplist_update(prop_data_t *data)
{
zfs_handle_t *root_hdl;
if ((root_hdl = zfs_open(data->lbh->lzh, data->lbh->root,
ZFS_TYPE_FILESYSTEM)) == NULL)
return (BE_ERR_ZFSOPEN);
/* XXX TODO: some error checking here */
zfs_iter_filesystems(root_hdl, prop_list_builder_cb, data);
zfs_close(root_hdl);
return (0);
}
static int
snapshot_proplist_update(zfs_handle_t *hdl, prop_data_t *data)
{
return (zfs_iter_snapshots_sorted(hdl, prop_list_builder_cb, data));
}
int
be_prop_list_alloc(nvlist_t **be_list)
{
return (nvlist_alloc(be_list, NV_UNIQUE_NAME, KM_SLEEP));
}
/*
* frees property list and its children
*/
void
be_prop_list_free(nvlist_t *be_list)
{
nvlist_t *prop_list;
nvpair_t *be_pair;
be_pair = nvlist_next_nvpair(be_list, NULL);
if (nvpair_value_nvlist(be_pair, &prop_list) == 0)
nvlist_free(prop_list);
while ((be_pair = nvlist_next_nvpair(be_list, be_pair)) != NULL) {
if (nvpair_value_nvlist(be_pair, &prop_list) == 0)
nvlist_free(prop_list);
}
}
/*
* Usage
*/
int
be_exists(libbe_handle_t *lbh, char *be)
{
char buf[BE_MAXPATHLEN];
nvlist_t *dsprops;
char *mntpoint;
bool valid;
be_root_concat(lbh, be, buf);
if (!zfs_dataset_exists(lbh->lzh, buf, ZFS_TYPE_DATASET))
return (BE_ERR_NOENT);
/* Also check if it's mounted at / */
if (be_prop_list_alloc(&dsprops) != 0)
return (BE_ERR_UNKNOWN);
if (be_get_dataset_props(lbh, buf, dsprops) != 0) {
nvlist_free(dsprops);
return (BE_ERR_UNKNOWN);
}
if (nvlist_lookup_string(dsprops, "mountpoint", &mntpoint) == 0) {
valid = (strcmp(mntpoint, "/") == 0);
nvlist_free(dsprops);
return (valid ? BE_ERR_SUCCESS : BE_ERR_BADMOUNT);
}
nvlist_free(dsprops);
return (BE_ERR_BADMOUNT);
}