journal: add object sanity check to journal_file_move_to_object()

Introduce journal_file_check_object(), which does lightweight object
sanity checks, and use it in journal_file_move_to_object(), so that we
will catch certain corrupted objects in the journal file.

This fixes #6447, where we had only partially written out OBJECT_ENTRY
(ObjectHeader written, but rest of object zero bytes), causing
"journalctl --list-boots" to fail.

  $ builddir.vanilla/journalctl --list-boots -D bug6447/
  Failed to determine boots: No data available

  $ builddir.patched/journalctl --list-boots -D bug6447/
  -52 22633da1c5374a728d6c215e2c301dc2 Mon 2017-07-10 05:29:21 EEST—Mon 2017-07-10 05:31:51 EEST
  -51 2253aab9ea7e4a2598f2abda82939eff Mon 2017-07-10 05:32:22 EEST—Mon 2017-07-10 05:36:49 EEST
  -50 ef0d85d35c74486fa4104f9d6391b6ba Mon 2017-07-10 05:40:33 EEST—Mon 2017-07-10 05:40:40 EEST
  [...]

Note that journal_file_check_object() is similar to
journal_file_object_verify(). The most expensive checks are omitted, as
they would slow down every journal_file_move_to_object() call too much.

With this implementation, the added overhead is small, for example when
dumping some journal content to /dev/null
(built with -Dbuildtype=debugoptimized -Db_ndebug=true):

 Performance counter stats for 'builddir.vanilla/journalctl -D 76f4d4c3406945f9a60d3ca8763aa754/':

      12542,311634      task-clock:u (msec)       #    1,000 CPUs utilized
                 0      context-switches:u        #    0,000 K/sec
                 0      cpu-migrations:u          #    0,000 K/sec
            80 100      page-faults:u             #    0,006 M/sec
    41 786 963 456      cycles:u                  #    3,332 GHz
   105 453 864 770      instructions:u            #    2,52  insn per cycle
    24 342 227 334      branches:u                # 1940,809 M/sec
       105 709 217      branch-misses:u           #    0,43% of all branches

      12,545199291 seconds time elapsed

 Performance counter stats for 'builddir.patched/journalctl -D 76f4d4c3406945f9a60d3ca8763aa754/':

      12734,723233      task-clock:u (msec)       #    1,000 CPUs utilized
                 0      context-switches:u        #    0,000 K/sec
                 0      cpu-migrations:u          #    0,000 K/sec
            80 693      page-faults:u             #    0,006 M/sec
    42 661 017 429      cycles:u                  #    3,350 GHz
   107 696 985 865      instructions:u            #    2,52  insn per cycle
    24 950 526 745      branches:u                # 1959,252 M/sec
       101 762 806      branch-misses:u           #    0,41% of all branches

      12,737527327 seconds time elapsed

Fixes #6447.
This commit is contained in:
Tommi Rantala 2017-09-19 11:10:49 +03:00
parent d82611c918
commit 24754f3694

View file

@ -770,6 +770,169 @@ static uint64_t minimum_header_size(Object *o) {
return table[o->object.type];
}
/* Lightweight object checks. We want this to be fast, so that we won't
* slowdown every journal_file_move_to_object() call too much. */
static int journal_file_check_object(JournalFile *f, uint64_t offset, Object *o) {
assert(f);
assert(o);
switch (o->object.type) {
case OBJECT_DATA: {
if ((le64toh(o->data.entry_offset) == 0) ^ (le64toh(o->data.n_entries) == 0)) {
log_debug("Bad n_entries: %"PRIu64": %"PRIu64,
o->data.n_entries, offset);
return -EBADMSG;
}
if (le64toh(o->object.size) - offsetof(DataObject, payload) <= 0) {
log_debug("Bad object size (<= %zu): %"PRIu64": %"PRIu64,
offsetof(DataObject, payload),
le64toh(o->object.size),
offset);
return -EBADMSG;
}
if (!VALID64(o->data.next_hash_offset) ||
!VALID64(o->data.next_field_offset) ||
!VALID64(o->data.entry_offset) ||
!VALID64(o->data.entry_array_offset)) {
log_debug("Invalid offset, next_hash_offset="OFSfmt", next_field_offset="OFSfmt
", entry_offset="OFSfmt", entry_array_offset="OFSfmt": %"PRIu64,
o->data.next_hash_offset,
o->data.next_field_offset,
o->data.entry_offset,
o->data.entry_array_offset,
offset);
return -EBADMSG;
}
break;
}
case OBJECT_FIELD:
if (le64toh(o->object.size) - offsetof(FieldObject, payload) <= 0) {
log_debug(
"Bad field size (<= %zu): %"PRIu64": %"PRIu64,
offsetof(FieldObject, payload),
le64toh(o->object.size),
offset);
return -EBADMSG;
}
if (!VALID64(o->field.next_hash_offset) ||
!VALID64(o->field.head_data_offset)) {
log_debug(
"Invalid offset, next_hash_offset="OFSfmt
", head_data_offset="OFSfmt": %"PRIu64,
o->field.next_hash_offset,
o->field.head_data_offset,
offset);
return -EBADMSG;
}
break;
case OBJECT_ENTRY:
if ((le64toh(o->object.size) - offsetof(EntryObject, items)) % sizeof(EntryItem) != 0) {
log_debug(
"Bad entry size (<= %zu): %"PRIu64": %"PRIu64,
offsetof(EntryObject, items),
le64toh(o->object.size),
offset);
return -EBADMSG;
}
if ((le64toh(o->object.size) - offsetof(EntryObject, items)) / sizeof(EntryItem) <= 0) {
log_debug(
"Invalid number items in entry: %"PRIu64": %"PRIu64,
(le64toh(o->object.size) - offsetof(EntryObject, items)) / sizeof(EntryItem),
offset);
return -EBADMSG;
}
if (le64toh(o->entry.seqnum) <= 0) {
log_debug(
"Invalid entry seqnum: %"PRIx64": %"PRIu64,
le64toh(o->entry.seqnum),
offset);
return -EBADMSG;
}
if (!VALID_REALTIME(le64toh(o->entry.realtime))) {
log_debug(
"Invalid entry realtime timestamp: %"PRIu64": %"PRIu64,
le64toh(o->entry.realtime),
offset);
return -EBADMSG;
}
if (!VALID_MONOTONIC(le64toh(o->entry.monotonic))) {
log_debug(
"Invalid entry monotonic timestamp: %"PRIu64": %"PRIu64,
le64toh(o->entry.monotonic),
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) {
log_debug(
"Invalid %s hash table size: %"PRIu64": %"PRIu64,
o->object.type == OBJECT_DATA_HASH_TABLE ? "data" : "field",
le64toh(o->object.size),
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) {
log_debug(
"Invalid object entry array size: %"PRIu64": %"PRIu64,
le64toh(o->object.size),
offset);
return -EBADMSG;
}
if (!VALID64(o->entry_array.next_entry_array_offset)) {
log_debug(
"Invalid object entry array next_entry_array_offset: "OFSfmt": %"PRIu64,
o->entry_array.next_entry_array_offset,
offset);
return -EBADMSG;
}
break;
case OBJECT_TAG:
if (le64toh(o->object.size) != sizeof(TagObject)) {
log_debug(
"Invalid object tag size: %"PRIu64": %"PRIu64,
le64toh(o->object.size),
offset);
return -EBADMSG;
}
if (!VALID_EPOCH(o->tag.epoch)) {
log_debug(
"Invalid object tag epoch: %"PRIu64": %"PRIu64,
o->tag.epoch,
offset);
return -EBADMSG;
}
break;
}
return 0;
}
int journal_file_move_to_object(JournalFile *f, ObjectType type, uint64_t offset, Object **ret) {
int r;
void *t;
@ -831,6 +994,10 @@ int journal_file_move_to_object(JournalFile *f, ObjectType type, uint64_t offset
o = (Object*) t;
}
r = journal_file_check_object(f, offset, o);
if (r < 0)
return r;
*ret = o;
return 0;
}