blob: 4d35168591e10d56dffab5acdd1173d82e5af5b0 [file] [log] [blame] [raw]
/* SPDX-License-Identifier: LGPL-2.1+ */
/***
This file is part of systemd.
Copyright 2012 Lennart Poettering
systemd 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.
systemd 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 systemd; If not, see <http://www.gnu.org/licenses/>.
***/
#include <fcntl.h>
#include <stddef.h>
#include <sys/mman.h>
#include <unistd.h>
#include "alloc-util.h"
#include "compress.h"
#include "fd-util.h"
#include "fileio.h"
#include "fs-util.h"
#include "journal-authenticate.h"
#include "journal-def.h"
#include "journal-file.h"
#include "journal-verify.h"
#include "lookup3.h"
#include "macro.h"
#include "terminal-util.h"
#include "util.h"
static void draw_progress(uint64_t p, usec_t *last_usec) {
unsigned n, i, j, k;
usec_t z, x;
if (!on_tty())
return;
z = now(CLOCK_MONOTONIC);
x = *last_usec;
if (x != 0 && x + 40 * USEC_PER_MSEC > z)
return;
*last_usec = z;
n = (3 * columns()) / 4;
j = (n * (unsigned) p) / 65535ULL;
k = n - j;
fputs("\r", stdout);
if (colors_enabled())
fputs("\x1B[?25l" ANSI_HIGHLIGHT_GREEN, stdout);
for (i = 0; i < j; i++)
fputs("\xe2\x96\x88", stdout);
fputs(ANSI_NORMAL, stdout);
for (i = 0; i < k; i++)
fputs("\xe2\x96\x91", stdout);
printf(" %3"PRIu64"%%", 100U * p / 65535U);
fputs("\r", stdout);
if (colors_enabled())
fputs("\x1B[?25h", stdout);
fflush(stdout);
}
static uint64_t scale_progress(uint64_t scale, uint64_t p, uint64_t m) {
/* Calculates scale * p / m, but handles m == 0 safely, and saturates */
if (p >= m || m == 0)
return scale;
return scale * p / m;
}
static void flush_progress(void) {
unsigned n, i;
if (!on_tty())
return;
n = (3 * columns()) / 4;
putchar('\r');
for (i = 0; i < n + 5; i++)
putchar(' ');
putchar('\r');
fflush(stdout);
}
#define debug(_offset, _fmt, ...) do { \
flush_progress(); \
log_debug(OFSfmt": " _fmt, _offset, ##__VA_ARGS__); \
} while (0)
#define warning(_offset, _fmt, ...) do { \
flush_progress(); \
log_warning(OFSfmt": " _fmt, _offset, ##__VA_ARGS__); \
} while (0)
#define error(_offset, _fmt, ...) do { \
flush_progress(); \
log_error(OFSfmt": " _fmt, (uint64_t)_offset, ##__VA_ARGS__); \
} while (0)
#define error_errno(_offset, error, _fmt, ...) do { \
flush_progress(); \
log_error_errno(error, OFSfmt": " _fmt, (uint64_t)_offset, ##__VA_ARGS__); \
} while (0)
static int journal_file_object_verify(JournalFile *f, uint64_t offset, Object *o) {
uint64_t i;
assert(f);
assert(offset);
assert(o);
/* This does various superficial tests about the length an
* possible field values. It does not follow any references to
* other objects. */
if ((o->object.flags & OBJECT_COMPRESSED_XZ) &&
o->object.type != OBJECT_DATA) {
error(offset, "Found compressed object that isn't of type DATA, which is not allowed.");
return -EBADMSG;
}
switch (o->object.type) {
case OBJECT_DATA: {
uint64_t h1, h2;
int compression, r;
if (le64toh(o->data.entry_offset) == 0)
warning(offset, "Unused data (entry_offset==0)");
if ((le64toh(o->data.entry_offset) == 0) ^ (le64toh(o->data.n_entries) == 0)) {
error(offset, "Bad n_entries: %"PRIu64, le64toh(o->data.n_entries));
return -EBADMSG;
}
if (le64toh(o->object.size) - offsetof(DataObject, payload) <= 0) {
error(offset, "Bad object size (<= %zu): %"PRIu64,
offsetof(DataObject, payload),
le64toh(o->object.size));
return -EBADMSG;
}
h1 = le64toh(o->data.hash);
compression = o->object.flags & OBJECT_COMPRESSION_MASK;
if (compression) {
_cleanup_free_ void *b = NULL;
size_t alloc = 0, b_size;
r = decompress_blob(compression,
o->data.payload,
le64toh(o->object.size) - offsetof(Object, data.payload),
&b, &alloc, &b_size, 0);
if (r < 0) {
error_errno(offset, r, "%s decompression failed: %m",
object_compressed_to_string(compression));
return r;
}
h2 = hash64(b, b_size);
} else
h2 = hash64(o->data.payload, le64toh(o->object.size) - offsetof(Object, data.payload));
if (h1 != h2) {
error(offset, "Invalid hash (%08"PRIx64" vs. %08"PRIx64, h1, h2);
return -EBADMSG;
}
if (!VALID64(le64toh(o->data.next_hash_offset)) ||
!VALID64(le64toh(o->data.next_field_offset)) ||
!VALID64(le64toh(o->data.entry_offset)) ||
!VALID64(le64toh(o->data.entry_array_offset))) {
error(offset, "Invalid offset (next_hash_offset="OFSfmt", next_field_offset="OFSfmt", entry_offset="OFSfmt", entry_array_offset="OFSfmt,
le64toh(o->data.next_hash_offset),
le64toh(o->data.next_field_offset),
le64toh(o->data.entry_offset),
le64toh(o->data.entry_array_offset));
return -EBADMSG;
}
break;
}
case OBJECT_FIELD:
if (le64toh(o->object.size) - offsetof(FieldObject, payload) <= 0) {
error(offset,
"Bad field size (<= %zu): %"PRIu64,
offsetof(FieldObject, payload),
le64toh(o->object.size));
return -EBADMSG;
}
if (!VALID64(le64toh(o->field.next_hash_offset)) ||
!VALID64(le64toh(o->field.head_data_offset))) {
error(offset,
"Invalid offset (next_hash_offset="OFSfmt", head_data_offset="OFSfmt,
le64toh(o->field.next_hash_offset),
le64toh(o->field.head_data_offset));
return -EBADMSG;
}
break;
case OBJECT_ENTRY:
if ((le64toh(o->object.size) - offsetof(EntryObject, items)) % sizeof(EntryItem) != 0) {
error(offset,
"Bad entry size (<= %zu): %"PRIu64,
offsetof(EntryObject, items),
le64toh(o->object.size));
return -EBADMSG;
}
if ((le64toh(o->object.size) - offsetof(EntryObject, items)) / sizeof(EntryItem) <= 0) {
error(offset,
"Invalid number items in entry: %"PRIu64,
(le64toh(o->object.size) - offsetof(EntryObject, items)) / sizeof(EntryItem));
return -EBADMSG;
}
if (le64toh(o->entry.seqnum) <= 0) {
error(offset,
"Invalid entry seqnum: %"PRIx64,
le64toh(o->entry.seqnum));
return -EBADMSG;
}
if (!VALID_REALTIME(le64toh(o->entry.realtime))) {
error(offset,
"Invalid entry realtime timestamp: %"PRIu64,
le64toh(o->entry.realtime));
return -EBADMSG;
}
if (!VALID_MONOTONIC(le64toh(o->entry.monotonic))) {
error(offset,
"Invalid entry monotonic timestamp: %"PRIu64,
le64toh(o->entry.monotonic));
return -EBADMSG;
}
for (i = 0; i < journal_file_entry_n_items(o); i++) {
if (le64toh(o->entry.items[i].object_offset) == 0 ||
!VALID64(le64toh(o->entry.items[i].object_offset))) {
error(offset,
"Invalid entry item (%"PRIu64"/%"PRIu64" offset: "OFSfmt,
i, journal_file_entry_n_items(o),
le64toh(o->entry.items[i].object_offset));
return -EBADMSG;
}
}
break;
case OBJECT_DATA_HASH_TABLE:
case OBJECT_FIELD_HASH_TABLE:
if ((le64toh(o->object.size) - offsetof(HashTableObject, items)) % sizeof(HashItem) != 0 ||
(le64toh(o->object.size) - offsetof(HashTableObject, items)) / sizeof(HashItem) <= 0) {
error(offset,
"Invalid %s hash table size: %"PRIu64,
o->object.type == OBJECT_DATA_HASH_TABLE ? "data" : "field",
le64toh(o->object.size));
return -EBADMSG;
}
for (i = 0; i < journal_file_hash_table_n_items(o); i++) {
if (o->hash_table.items[i].head_hash_offset != 0 &&
!VALID64(le64toh(o->hash_table.items[i].head_hash_offset))) {
error(offset,
"Invalid %s hash table item (%"PRIu64"/%"PRIu64") head_hash_offset: "OFSfmt,
o->object.type == OBJECT_DATA_HASH_TABLE ? "data" : "field",
i, journal_file_hash_table_n_items(o),
le64toh(o->hash_table.items[i].head_hash_offset));
return -EBADMSG;
}
if (o->hash_table.items[i].tail_hash_offset != 0 &&
!VALID64(le64toh(o->hash_table.items[i].tail_hash_offset))) {
error(offset,
"Invalid %s hash table item (%"PRIu64"/%"PRIu64") tail_hash_offset: "OFSfmt,
o->object.type == OBJECT_DATA_HASH_TABLE ? "data" : "field",
i, journal_file_hash_table_n_items(o),
le64toh(o->hash_table.items[i].tail_hash_offset));
return -EBADMSG;
}
if ((o->hash_table.items[i].head_hash_offset != 0) !=
(o->hash_table.items[i].tail_hash_offset != 0)) {
error(offset,
"Invalid %s hash table item (%"PRIu64"/%"PRIu64"): head_hash_offset="OFSfmt" tail_hash_offset="OFSfmt,
o->object.type == OBJECT_DATA_HASH_TABLE ? "data" : "field",
i, journal_file_hash_table_n_items(o),
le64toh(o->hash_table.items[i].head_hash_offset),
le64toh(o->hash_table.items[i].tail_hash_offset));
return -EBADMSG;
}
}
break;
case OBJECT_ENTRY_ARRAY:
if ((le64toh(o->object.size) - offsetof(EntryArrayObject, items)) % sizeof(le64_t) != 0 ||
(le64toh(o->object.size) - offsetof(EntryArrayObject, items)) / sizeof(le64_t) <= 0) {
error(offset,
"Invalid object entry array size: %"PRIu64,
le64toh(o->object.size));
return -EBADMSG;
}
if (!VALID64(le64toh(o->entry_array.next_entry_array_offset))) {
error(offset,
"Invalid object entry array next_entry_array_offset: "OFSfmt,
le64toh(o->entry_array.next_entry_array_offset));
return -EBADMSG;
}
for (i = 0; i < journal_file_entry_array_n_items(o); i++)
if (le64toh(o->entry_array.items[i]) != 0 &&
!VALID64(le64toh(o->entry_array.items[i]))) {
error(offset,
"Invalid object entry array item (%"PRIu64"/%"PRIu64"): "OFSfmt,
i, journal_file_entry_array_n_items(o),
le64toh(o->entry_array.items[i]));
return -EBADMSG;
}
break;
case OBJECT_TAG:
if (le64toh(o->object.size) != sizeof(TagObject)) {
error(offset,
"Invalid object tag size: %"PRIu64,
le64toh(o->object.size));
return -EBADMSG;
}
if (!VALID_EPOCH(le64toh(o->tag.epoch))) {
error(offset,
"Invalid object tag epoch: %"PRIu64,
le64toh(o->tag.epoch));
return -EBADMSG;
}
break;
}
return 0;
}
static int write_uint64(int fd, uint64_t p) {
ssize_t k;
k = write(fd, &p, sizeof(p));
if (k < 0)
return -errno;
if (k != sizeof(p))
return -EIO;
return 0;
}
static int contains_uint64(MMapCache *m, MMapFileDescriptor *f, uint64_t n, uint64_t p) {
uint64_t a, b;
int r;
assert(m);
assert(f);
/* Bisection ... */
a = 0; b = n;
while (a < b) {
uint64_t c, *z;
c = (a + b) / 2;
r = mmap_cache_get(m, f, PROT_READ|PROT_WRITE, 0, false, c * sizeof(uint64_t), sizeof(uint64_t), NULL, (void **) &z, NULL);
if (r < 0)
return r;
if (*z == p)
return 1;
if (a + 1 >= b)
return 0;
if (p < *z)
b = c;
else
a = c;
}
return 0;
}
static int entry_points_to_data(
JournalFile *f,
MMapFileDescriptor *cache_entry_fd,
uint64_t n_entries,
uint64_t entry_p,
uint64_t data_p) {
int r;
uint64_t i, n, a;
Object *o;
bool found = false;
assert(f);
assert(cache_entry_fd);
if (!contains_uint64(f->mmap, cache_entry_fd, n_entries, entry_p)) {
error(data_p, "Data object references invalid entry at "OFSfmt, entry_p);
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_ENTRY, entry_p, &o);
if (r < 0)
return r;
n = journal_file_entry_n_items(o);
for (i = 0; i < n; i++)
if (le64toh(o->entry.items[i].object_offset) == data_p) {
found = true;
break;
}
if (!found) {
error(entry_p, "Data object at "OFSfmt" not referenced by linked entry", data_p);
return -EBADMSG;
}
/* Check if this entry is also in main entry array. Since the
* main entry array has already been verified we can rely on
* its consistency. */
i = 0;
n = le64toh(f->header->n_entries);
a = le64toh(f->header->entry_array_offset);
while (i < n) {
uint64_t m, u;
r = journal_file_move_to_object(f, OBJECT_ENTRY_ARRAY, a, &o);
if (r < 0)
return r;
m = journal_file_entry_array_n_items(o);
u = MIN(n - i, m);
if (entry_p <= le64toh(o->entry_array.items[u-1])) {
uint64_t x, y, z;
x = 0;
y = u;
while (x < y) {
z = (x + y) / 2;
if (le64toh(o->entry_array.items[z]) == entry_p)
return 0;
if (x + 1 >= y)
break;
if (entry_p < le64toh(o->entry_array.items[z]))
y = z;
else
x = z;
}
error(entry_p, "Entry object doesn't exist in main entry array");
return -EBADMSG;
}
i += u;
a = le64toh(o->entry_array.next_entry_array_offset);
}
return 0;
}
static int verify_data(
JournalFile *f,
Object *o, uint64_t p,
MMapFileDescriptor *cache_entry_fd, uint64_t n_entries,
MMapFileDescriptor *cache_entry_array_fd, uint64_t n_entry_arrays) {
uint64_t i, n, a, last, q;
int r;
assert(f);
assert(o);
assert(cache_entry_fd);
assert(cache_entry_array_fd);
n = le64toh(o->data.n_entries);
a = le64toh(o->data.entry_array_offset);
/* Entry array means at least two objects */
if (a && n < 2) {
error(p, "Entry array present (entry_array_offset="OFSfmt", but n_entries=%"PRIu64")", a, n);
return -EBADMSG;
}
if (n == 0)
return 0;
/* We already checked that earlier */
assert(o->data.entry_offset);
last = q = le64toh(o->data.entry_offset);
r = entry_points_to_data(f, cache_entry_fd, n_entries, q, p);
if (r < 0)
return r;
i = 1;
while (i < n) {
uint64_t next, m, j;
if (a == 0) {
error(p, "Array chain too short");
return -EBADMSG;
}
if (!contains_uint64(f->mmap, cache_entry_array_fd, n_entry_arrays, a)) {
error(p, "Invalid array offset "OFSfmt, a);
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_ENTRY_ARRAY, a, &o);
if (r < 0)
return r;
next = le64toh(o->entry_array.next_entry_array_offset);
if (next != 0 && next <= a) {
error(p, "Array chain has cycle (jumps back from "OFSfmt" to "OFSfmt")", a, next);
return -EBADMSG;
}
m = journal_file_entry_array_n_items(o);
for (j = 0; i < n && j < m; i++, j++) {
q = le64toh(o->entry_array.items[j]);
if (q <= last) {
error(p, "Data object's entry array not sorted");
return -EBADMSG;
}
last = q;
r = entry_points_to_data(f, cache_entry_fd, n_entries, q, p);
if (r < 0)
return r;
/* Pointer might have moved, reposition */
r = journal_file_move_to_object(f, OBJECT_ENTRY_ARRAY, a, &o);
if (r < 0)
return r;
}
a = next;
}
return 0;
}
static int verify_hash_table(
JournalFile *f,
MMapFileDescriptor *cache_data_fd, uint64_t n_data,
MMapFileDescriptor *cache_entry_fd, uint64_t n_entries,
MMapFileDescriptor *cache_entry_array_fd, uint64_t n_entry_arrays,
usec_t *last_usec,
bool show_progress) {
uint64_t i, n;
int r;
assert(f);
assert(cache_data_fd);
assert(cache_entry_fd);
assert(cache_entry_array_fd);
assert(last_usec);
n = le64toh(f->header->data_hash_table_size) / sizeof(HashItem);
if (n <= 0)
return 0;
r = journal_file_map_data_hash_table(f);
if (r < 0)
return log_error_errno(r, "Failed to map data hash table: %m");
for (i = 0; i < n; i++) {
uint64_t last = 0, p;
if (show_progress)
draw_progress(0xC000 + scale_progress(0x3FFF, i, n), last_usec);
p = le64toh(f->data_hash_table[i].head_hash_offset);
while (p != 0) {
Object *o;
uint64_t next;
if (!contains_uint64(f->mmap, cache_data_fd, n_data, p)) {
error(p, "Invalid data object at hash entry %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_DATA, p, &o);
if (r < 0)
return r;
next = le64toh(o->data.next_hash_offset);
if (next != 0 && next <= p) {
error(p, "Hash chain has a cycle in hash entry %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
if (le64toh(o->data.hash) % n != i) {
error(p, "Hash value mismatch in hash entry %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
r = verify_data(f, o, p, cache_entry_fd, n_entries, cache_entry_array_fd, n_entry_arrays);
if (r < 0)
return r;
last = p;
p = next;
}
if (last != le64toh(f->data_hash_table[i].tail_hash_offset)) {
error(p, "Tail hash pointer mismatch in hash table");
return -EBADMSG;
}
}
return 0;
}
static int data_object_in_hash_table(JournalFile *f, uint64_t hash, uint64_t p) {
uint64_t n, h, q;
int r;
assert(f);
n = le64toh(f->header->data_hash_table_size) / sizeof(HashItem);
if (n <= 0)
return 0;
r = journal_file_map_data_hash_table(f);
if (r < 0)
return log_error_errno(r, "Failed to map data hash table: %m");
h = hash % n;
q = le64toh(f->data_hash_table[h].head_hash_offset);
while (q != 0) {
Object *o;
if (p == q)
return 1;
r = journal_file_move_to_object(f, OBJECT_DATA, q, &o);
if (r < 0)
return r;
q = le64toh(o->data.next_hash_offset);
}
return 0;
}
static int verify_entry(
JournalFile *f,
Object *o, uint64_t p,
MMapFileDescriptor *cache_data_fd, uint64_t n_data) {
uint64_t i, n;
int r;
assert(f);
assert(o);
assert(cache_data_fd);
n = journal_file_entry_n_items(o);
for (i = 0; i < n; i++) {
uint64_t q, h;
Object *u;
q = le64toh(o->entry.items[i].object_offset);
h = le64toh(o->entry.items[i].hash);
if (!contains_uint64(f->mmap, cache_data_fd, n_data, q)) {
error(p, "Invalid data object of entry");
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_DATA, q, &u);
if (r < 0)
return r;
if (le64toh(u->data.hash) != h) {
error(p, "Hash mismatch for data object of entry");
return -EBADMSG;
}
r = data_object_in_hash_table(f, h, q);
if (r < 0)
return r;
if (r == 0) {
error(p, "Data object missing from hash table");
return -EBADMSG;
}
}
return 0;
}
static int verify_entry_array(
JournalFile *f,
MMapFileDescriptor *cache_data_fd, uint64_t n_data,
MMapFileDescriptor *cache_entry_fd, uint64_t n_entries,
MMapFileDescriptor *cache_entry_array_fd, uint64_t n_entry_arrays,
usec_t *last_usec,
bool show_progress) {
uint64_t i = 0, a, n, last = 0;
int r;
assert(f);
assert(cache_data_fd);
assert(cache_entry_fd);
assert(cache_entry_array_fd);
assert(last_usec);
n = le64toh(f->header->n_entries);
a = le64toh(f->header->entry_array_offset);
while (i < n) {
uint64_t next, m, j;
Object *o;
if (show_progress)
draw_progress(0x8000 + scale_progress(0x3FFF, i, n), last_usec);
if (a == 0) {
error(a, "Array chain too short at %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
if (!contains_uint64(f->mmap, cache_entry_array_fd, n_entry_arrays, a)) {
error(a, "Invalid array %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_ENTRY_ARRAY, a, &o);
if (r < 0)
return r;
next = le64toh(o->entry_array.next_entry_array_offset);
if (next != 0 && next <= a) {
error(a, "Array chain has cycle at %"PRIu64" of %"PRIu64" (jumps back from to "OFSfmt")", i, n, next);
return -EBADMSG;
}
m = journal_file_entry_array_n_items(o);
for (j = 0; i < n && j < m; i++, j++) {
uint64_t p;
p = le64toh(o->entry_array.items[j]);
if (p <= last) {
error(a, "Entry array not sorted at %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
last = p;
if (!contains_uint64(f->mmap, cache_entry_fd, n_entries, p)) {
error(a, "Invalid array entry at %"PRIu64" of %"PRIu64, i, n);
return -EBADMSG;
}
r = journal_file_move_to_object(f, OBJECT_ENTRY, p, &o);
if (r < 0)
return r;
r = verify_entry(f, o, p, cache_data_fd, n_data);
if (r < 0)
return r;
/* Pointer might have moved, reposition */
r = journal_file_move_to_object(f, OBJECT_ENTRY_ARRAY, a, &o);
if (r < 0)
return r;
}
a = next;
}
return 0;
}
int journal_file_verify(
JournalFile *f,
const char *key,
usec_t *first_contained, usec_t *last_validated, usec_t *last_contained,
bool show_progress) {
int r;
Object *o;
uint64_t p = 0, last_epoch = 0, last_tag_realtime = 0, last_sealed_realtime = 0;
uint64_t entry_seqnum = 0, entry_monotonic = 0, entry_realtime = 0;
sd_id128_t entry_boot_id;
bool entry_seqnum_set = false, entry_monotonic_set = false, entry_realtime_set = false, found_main_entry_array = false;
uint64_t n_weird = 0, n_objects = 0, n_entries = 0, n_data = 0, n_fields = 0, n_data_hash_tables = 0, n_field_hash_tables = 0, n_entry_arrays = 0, n_tags = 0;
usec_t last_usec = 0;
int data_fd = -1, entry_fd = -1, entry_array_fd = -1;
MMapFileDescriptor *cache_data_fd = NULL, *cache_entry_fd = NULL, *cache_entry_array_fd = NULL;
unsigned i;
bool found_last = false;
const char *tmp_dir = NULL;
#if HAVE_GCRYPT
uint64_t last_tag = 0;
#endif
assert(f);
if (key) {
#if HAVE_GCRYPT
r = journal_file_parse_verification_key(f, key);
if (r < 0) {
log_error("Failed to parse seed.");
return r;
}
#else
return -EOPNOTSUPP;
#endif
} else if (f->seal)
return -ENOKEY;
r = var_tmp_dir(&tmp_dir);
if (r < 0) {
log_error_errno(r, "Failed to determine temporary directory: %m");
goto fail;
}
data_fd = open_tmpfile_unlinkable(tmp_dir, O_RDWR | O_CLOEXEC);
if (data_fd < 0) {
r = log_error_errno(data_fd, "Failed to create data file: %m");
goto fail;
}
entry_fd = open_tmpfile_unlinkable(tmp_dir, O_RDWR | O_CLOEXEC);
if (entry_fd < 0) {
r = log_error_errno(entry_fd, "Failed to create entry file: %m");
goto fail;
}
entry_array_fd = open_tmpfile_unlinkable(tmp_dir, O_RDWR | O_CLOEXEC);
if (entry_array_fd < 0) {
r = log_error_errno(entry_array_fd,
"Failed to create entry array file: %m");
goto fail;
}
cache_data_fd = mmap_cache_add_fd(f->mmap, data_fd);
if (!cache_data_fd) {
r = log_oom();
goto fail;
}
cache_entry_fd = mmap_cache_add_fd(f->mmap, entry_fd);
if (!cache_entry_fd) {
r = log_oom();
goto fail;
}
cache_entry_array_fd = mmap_cache_add_fd(f->mmap, entry_array_fd);
if (!cache_entry_array_fd) {
r = log_oom();
goto fail;
}
if (le32toh(f->header->compatible_flags) & ~HEADER_COMPATIBLE_SUPPORTED) {
log_error("Cannot verify file with unknown extensions.");
r = -EOPNOTSUPP;
goto fail;
}
for (i = 0; i < sizeof(f->header->reserved); i++)
if (f->header->reserved[i] != 0) {
error(offsetof(Header, reserved[i]), "Reserved field is non-zero");
r = -EBADMSG;
goto fail;
}
/* First iteration: we go through all objects, verify the
* superficial structure, headers, hashes. */
p = le64toh(f->header->header_size);
for (;;) {
/* Early exit if there are no objects in the file, at all */
if (le64toh(f->header->tail_object_offset) == 0)
break;
if (show_progress)
draw_progress(scale_progress(0x7FFF, p, le64toh(f->header->tail_object_offset)), &last_usec);
r = journal_file_move_to_object(f, OBJECT_UNUSED, p, &o);
if (r < 0) {
error(p, "Invalid object");
goto fail;
}
if (p > le64toh(f->header->tail_object_offset)) {
error(offsetof(Header, tail_object_offset), "Invalid tail object pointer");
r = -EBADMSG;
goto fail;
}
n_objects++;
r = journal_file_object_verify(f, p, o);
if (r < 0) {
error_errno(p, r, "Invalid object contents: %m");
goto fail;
}
if ((o->object.flags & OBJECT_COMPRESSED_XZ) &&
(o->object.flags & OBJECT_COMPRESSED_LZ4)) {
error(p, "Objected with double compression");
r = -EINVAL;
goto fail;
}
if ((o->object.flags & OBJECT_COMPRESSED_XZ) && !JOURNAL_HEADER_COMPRESSED_XZ(f->header)) {
error(p, "XZ compressed object in file without XZ compression");
r = -EBADMSG;
goto fail;
}
if ((o->object.flags & OBJECT_COMPRESSED_LZ4) && !JOURNAL_HEADER_COMPRESSED_LZ4(f->header)) {
error(p, "LZ4 compressed object in file without LZ4 compression");
r = -EBADMSG;
goto fail;
}
switch (o->object.type) {
case OBJECT_DATA:
r = write_uint64(data_fd, p);
if (r < 0)
goto fail;
n_data++;
break;
case OBJECT_FIELD:
n_fields++;
break;
case OBJECT_ENTRY:
if (JOURNAL_HEADER_SEALED(f->header) && n_tags <= 0) {
error(p, "First entry before first tag");
r = -EBADMSG;
goto fail;
}
r = write_uint64(entry_fd, p);
if (r < 0)
goto fail;
if (le64toh(o->entry.realtime) < last_tag_realtime) {
error(p, "Older entry after newer tag");
r = -EBADMSG;
goto fail;
}
if (!entry_seqnum_set &&
le64toh(o->entry.seqnum) != le64toh(f->header->head_entry_seqnum)) {
error(p, "Head entry sequence number incorrect");
r = -EBADMSG;
goto fail;
}
if (entry_seqnum_set &&
entry_seqnum >= le64toh(o->entry.seqnum)) {
error(p, "Entry sequence number out of synchronization");
r = -EBADMSG;
goto fail;
}
entry_seqnum = le64toh(o->entry.seqnum);
entry_seqnum_set = true;
if (entry_monotonic_set &&
sd_id128_equal(entry_boot_id, o->entry.boot_id) &&
entry_monotonic > le64toh(o->entry.monotonic)) {
error(p, "Entry timestamp out of synchronization");
r = -EBADMSG;
goto fail;
}
entry_monotonic = le64toh(o->entry.monotonic);
entry_boot_id = o->entry.boot_id;
entry_monotonic_set = true;
if (!entry_realtime_set &&
le64toh(o->entry.realtime) != le64toh(f->header->head_entry_realtime)) {
error(p, "Head entry realtime timestamp incorrect");
r = -EBADMSG;
goto fail;
}
entry_realtime = le64toh(o->entry.realtime);
entry_realtime_set = true;
n_entries++;
break;
case OBJECT_DATA_HASH_TABLE:
if (n_data_hash_tables > 1) {
error(p, "More than one data hash table");
r = -EBADMSG;
goto fail;
}
if (le64toh(f->header->data_hash_table_offset) != p + offsetof(HashTableObject, items) ||
le64toh(f->header->data_hash_table_size) != le64toh(o->object.size) - offsetof(HashTableObject, items)) {
error(p, "header fields for data hash table invalid");
r = -EBADMSG;
goto fail;
}
n_data_hash_tables++;
break;
case OBJECT_FIELD_HASH_TABLE:
if (n_field_hash_tables > 1) {
error(p, "More than one field hash table");
r = -EBADMSG;
goto fail;
}
if (le64toh(f->header->field_hash_table_offset) != p + offsetof(HashTableObject, items) ||
le64toh(f->header->field_hash_table_size) != le64toh(o->object.size) - offsetof(HashTableObject, items)) {
error(p, "Header fields for field hash table invalid");
r = -EBADMSG;
goto fail;
}
n_field_hash_tables++;
break;
case OBJECT_ENTRY_ARRAY:
r = write_uint64(entry_array_fd, p);
if (r < 0)
goto fail;
if (p == le64toh(f->header->entry_array_offset)) {
if (found_main_entry_array) {
error(p, "More than one main entry array");
r = -EBADMSG;
goto fail;
}
found_main_entry_array = true;
}
n_entry_arrays++;
break;
case OBJECT_TAG:
if (!JOURNAL_HEADER_SEALED(f->header)) {
error(p, "Tag object in file without sealing");
r = -EBADMSG;
goto fail;
}
if (le64toh(o->tag.seqnum) != n_tags + 1) {
error(p, "Tag sequence number out of synchronization");
r = -EBADMSG;
goto fail;
}
if (le64toh(o->tag.epoch) < last_epoch) {
error(p, "Epoch sequence out of synchronization");
r = -EBADMSG;
goto fail;
}
#if HAVE_GCRYPT
if (f->seal) {
uint64_t q, rt;
debug(p, "Checking tag %"PRIu64"...", le64toh(o->tag.seqnum));
rt = f->fss_start_usec + le64toh(o->tag.epoch) * f->fss_interval_usec;
if (entry_realtime_set && entry_realtime >= rt + f->fss_interval_usec) {
error(p, "tag/entry realtime timestamp out of synchronization");
r = -EBADMSG;
goto fail;
}
/* OK, now we know the epoch. So let's now set
* it, and calculate the HMAC for everything
* since the last tag. */
r = journal_file_fsprg_seek(f, le64toh(o->tag.epoch));
if (r < 0)
goto fail;
r = journal_file_hmac_start(f);
if (r < 0)
goto fail;
if (last_tag == 0) {
r = journal_file_hmac_put_header(f);
if (r < 0)
goto fail;
q = le64toh(f->header->header_size);
} else
q = last_tag;
while (q <= p) {
r = journal_file_move_to_object(f, OBJECT_UNUSED, q, &o);
if (r < 0)
goto fail;
r = journal_file_hmac_put_object(f, OBJECT_UNUSED, o, q);
if (r < 0)
goto fail;
q = q + ALIGN64(le64toh(o->object.size));
}
/* Position might have changed, let's reposition things */
r = journal_file_move_to_object(f, OBJECT_UNUSED, p, &o);
if (r < 0)
goto fail;
if (memcmp(o->tag.tag, gcry_md_read(f->hmac, 0), TAG_LENGTH) != 0) {
error(p, "Tag failed verification");
r = -EBADMSG;
goto fail;
}
f->hmac_running = false;
last_tag_realtime = rt;
last_sealed_realtime = entry_realtime;
}
last_tag = p + ALIGN64(le64toh(o->object.size));
#endif
last_epoch = le64toh(o->tag.epoch);
n_tags++;
break;
default:
n_weird++;
}
if (p == le64toh(f->header->tail_object_offset)) {
found_last = true;
break;
}
p = p + ALIGN64(le64toh(o->object.size));
};
if (!found_last && le64toh(f->header->tail_object_offset) != 0) {
error(le64toh(f->header->tail_object_offset), "Tail object pointer dead");
r = -EBADMSG;
goto fail;
}
if (n_objects != le64toh(f->header->n_objects)) {
error(offsetof(Header, n_objects), "Object number mismatch");
r = -EBADMSG;
goto fail;
}
if (n_entries != le64toh(f->header->n_entries)) {
error(offsetof(Header, n_entries), "Entry number mismatch");
r = -EBADMSG;
goto fail;
}
if (JOURNAL_HEADER_CONTAINS(f->header, n_data) &&
n_data != le64toh(f->header->n_data)) {
error(offsetof(Header, n_data), "Data number mismatch");
r = -EBADMSG;
goto fail;
}
if (JOURNAL_HEADER_CONTAINS(f->header, n_fields) &&
n_fields != le64toh(f->header->n_fields)) {
error(offsetof(Header, n_fields), "Field number mismatch");
r = -EBADMSG;
goto fail;
}
if (JOURNAL_HEADER_CONTAINS(f->header, n_tags) &&
n_tags != le64toh(f->header->n_tags)) {
error(offsetof(Header, n_tags), "Tag number mismatch");
r = -EBADMSG;
goto fail;
}
if (JOURNAL_HEADER_CONTAINS(f->header, n_entry_arrays) &&
n_entry_arrays != le64toh(f->header->n_entry_arrays)) {
error(offsetof(Header, n_entry_arrays), "Entry array number mismatch");
r = -EBADMSG;
goto fail;
}
if (!found_main_entry_array && le64toh(f->header->entry_array_offset) != 0) {
error(0, "Missing entry array");
r = -EBADMSG;
goto fail;
}
if (entry_seqnum_set &&
entry_seqnum != le64toh(f->header->tail_entry_seqnum)) {
error(offsetof(Header, tail_entry_seqnum), "Invalid tail seqnum");
r = -EBADMSG;
goto fail;
}
if (entry_monotonic_set &&
(sd_id128_equal(entry_boot_id, f->header->boot_id) &&
entry_monotonic != le64toh(f->header->tail_entry_monotonic))) {
error(0, "Invalid tail monotonic timestamp");
r = -EBADMSG;
goto fail;
}
if (entry_realtime_set && entry_realtime != le64toh(f->header->tail_entry_realtime)) {
error(0, "Invalid tail realtime timestamp");
r = -EBADMSG;
goto fail;
}
/* Second iteration: we follow all objects referenced from the
* two entry points: the object hash table and the entry
* array. We also check that everything referenced (directly
* or indirectly) in the data hash table also exists in the
* entry array, and vice versa. Note that we do not care for
* unreferenced objects. We only care that everything that is
* referenced is consistent. */
r = verify_entry_array(f,
cache_data_fd, n_data,
cache_entry_fd, n_entries,
cache_entry_array_fd, n_entry_arrays,
&last_usec,
show_progress);
if (r < 0)
goto fail;
r = verify_hash_table(f,
cache_data_fd, n_data,
cache_entry_fd, n_entries,
cache_entry_array_fd, n_entry_arrays,
&last_usec,
show_progress);
if (r < 0)
goto fail;
if (show_progress)
flush_progress();
mmap_cache_free_fd(f->mmap, cache_data_fd);
mmap_cache_free_fd(f->mmap, cache_entry_fd);
mmap_cache_free_fd(f->mmap, cache_entry_array_fd);
safe_close(data_fd);
safe_close(entry_fd);
safe_close(entry_array_fd);
if (first_contained)
*first_contained = le64toh(f->header->head_entry_realtime);
if (last_validated)
*last_validated = last_sealed_realtime;
if (last_contained)
*last_contained = le64toh(f->header->tail_entry_realtime);
return 0;
fail:
if (show_progress)
flush_progress();
log_error("File corruption detected at %s:"OFSfmt" (of %llu bytes, %"PRIu64"%%).",
f->path,
p,
(unsigned long long) f->last_stat.st_size,
100 * p / f->last_stat.st_size);
if (data_fd >= 0)
safe_close(data_fd);
if (entry_fd >= 0)
safe_close(entry_fd);
if (entry_array_fd >= 0)
safe_close(entry_array_fd);
if (cache_data_fd)
mmap_cache_free_fd(f->mmap, cache_data_fd);
if (cache_entry_fd)
mmap_cache_free_fd(f->mmap, cache_entry_fd);
if (cache_entry_array_fd)
mmap_cache_free_fd(f->mmap, cache_entry_array_fd);
return r;
}