/*
* set an chinese word into bitmap
* normarlly called mannny times on process initialize
* parameters:
* bm :bitmap which ts set into
* ts :an chinese word, e.g. 天气
*/
void word_set(char *bm, char *ts)
{
unsigned int pos = 0;
if (!ts || strlen(ts) > MAX_PHRASE_S_LEN) return;
BITMAP_POS(ts, strlen(ts), pos);
//char *s = ts; int len = strlen(ts);
if (BITMAP_GET(bm, pos)) {
//printf("%x seted\n", pos);
}
//if (pos == 0x24e1373a) printf("%s seted\n", ts);
BITMAP_SET(bm, pos);
}
/* Allocate a single block, first allocating an hbin (page) at the end
* of the current file if necessary. NB. To keep the implementation
* simple and more likely to be correct, we do not reuse existing free
* blocks.
*
* seg_len is the size of the block (this INCLUDES the block header).
* The header of the block is initialized to -seg_len (negative to
* indicate used). id[2] is the block ID (type), eg. "nk" for nk-
* record. The block bitmap is updated to show this block as valid.
* The rest of the contents of the block will be zero.
*
* **NB** Because allocate_block may reallocate the memory, all
* pointers into the memory become potentially invalid. I really
* love writing in C, can't you tell?
*
* Returns:
* > 0 : offset of new block
* 0 : error (errno set)
*/
static size_t
allocate_block (hive_h *h, size_t seg_len, const char id[2])
{
CHECK_WRITABLE (0);
if (seg_len < 4) {
/* The caller probably forgot to include the header. Note that
* value lists have no ID field, so seg_len == 4 would be possible
* for them, albeit unusual.
*/
SET_ERRNO (ERANGE, "refusing too small allocation (%zu)", seg_len);
return 0;
}
/* Refuse really large allocations. */
if (seg_len > HIVEX_MAX_ALLOCATION) {
SET_ERRNO (ERANGE, "refusing too large allocation (%zu)", seg_len);
return 0;
}
/* Round up allocation to multiple of 8 bytes. All blocks must be
* on an 8 byte boundary.
*/
seg_len = (seg_len + 7) & ~7;
/* Allocate a new page if necessary. */
if (h->endblocks == 0 || h->endblocks + seg_len > h->endpages) {
size_t newendblocks = allocate_page (h, seg_len);
if (newendblocks == 0)
return 0;
h->endblocks = newendblocks;
}
size_t offset = h->endblocks;
DEBUG (2, "new block at 0x%zx, size %zu", offset, seg_len);
struct ntreg_hbin_block *blockhdr =
(struct ntreg_hbin_block *) ((char *) h->addr + offset);
memset (blockhdr, 0, seg_len);
blockhdr->seg_len = htole32 (- (int32_t) seg_len);
if (id[0] && id[1] && seg_len >= sizeof (struct ntreg_hbin_block)) {
blockhdr->id[0] = id[0];
blockhdr->id[1] = id[1];
}
BITMAP_SET (h->bitmap, offset);
h->endblocks += seg_len;
/* If there is space after the last block in the last page, then we
* have to put a dummy free block header here to mark the rest of
* the page as free.
*/
ssize_t rem = h->endpages - h->endblocks;
if (rem > 0) {
DEBUG (2, "marking remainder of page free"
" starting at 0x%zx, size %zd", h->endblocks, rem);
assert (rem >= 4);
blockhdr = (struct ntreg_hbin_block *) ((char *) h->addr + h->endblocks);
blockhdr->seg_len = htole32 ((int32_t) rem);
}
return offset;
}
hive_h *
hivex_open (const char *filename, int flags)
{
hive_h *h = NULL;
assert (sizeof (struct ntreg_header) == 0x1000);
assert (offsetof (struct ntreg_header, csum) == 0x1fc);
h = calloc (1, sizeof *h);
if (h == NULL)
goto error;
h->msglvl = flags & HIVEX_OPEN_MSGLVL_MASK;
const char *debug = getenv ("HIVEX_DEBUG");
if (debug && STREQ (debug, "1"))
h->msglvl = 2;
DEBUG (2, "created handle %p", h);
h->writable = !!(flags & HIVEX_OPEN_WRITE);
h->unsafe = !!(flags & HIVEX_OPEN_UNSAFE);
h->filename = strdup (filename);
if (h->filename == NULL)
goto error;
#ifdef O_CLOEXEC
h->fd = open (filename, O_RDONLY | O_CLOEXEC | O_BINARY);
#else
h->fd = open (filename, O_RDONLY | O_BINARY);
#endif
if (h->fd == -1)
goto error;
#ifndef O_CLOEXEC
fcntl (h->fd, F_SETFD, FD_CLOEXEC);
#endif
struct stat statbuf;
if (fstat (h->fd, &statbuf) == -1)
goto error;
h->size = statbuf.st_size;
if (h->size < 0x2000) {
SET_ERRNO (EINVAL,
"%s: file is too small to be a Windows NT Registry hive file",
filename);
goto error;
}
if (!h->writable) {
h->addr = mmap (NULL, h->size, PROT_READ, MAP_SHARED, h->fd, 0);
if (h->addr == MAP_FAILED)
goto error;
DEBUG (2, "mapped file at %p", h->addr);
} else {
h->addr = malloc (h->size);
if (h->addr == NULL)
goto error;
if (full_read (h->fd, h->addr, h->size) < h->size)
goto error;
/* We don't need the file descriptor along this path, since we
* have read all the data.
*/
if (close (h->fd) == -1)
goto error;
h->fd = -1;
}
/* Check header. */
if (h->hdr->magic[0] != 'r' ||
h->hdr->magic[1] != 'e' ||
h->hdr->magic[2] != 'g' ||
h->hdr->magic[3] != 'f') {
SET_ERRNO (ENOTSUP,
"%s: not a Windows NT Registry hive file", filename);
goto error;
}
/* Check major version. */
uint32_t major_ver = le32toh (h->hdr->major_ver);
if (major_ver != 1) {
SET_ERRNO (ENOTSUP,
"%s: hive file major version %" PRIu32 " (expected 1)",
filename, major_ver);
goto error;
}
h->bitmap = calloc (1 + h->size / 32, 1);
if (h->bitmap == NULL)
goto error;
/* Header checksum. */
uint32_t sum = header_checksum (h);
if (sum != le32toh (h->hdr->csum)) {
SET_ERRNO (EINVAL, "%s: bad checksum in hive header", filename);
goto error;
}
for (int t=0; t<nr_recode_types; t++) {
gl_lock_init (h->iconv_cache[t].mutex);
h->iconv_cache[t].handle = NULL;
}
/* Last modified time. */
h->last_modified = le64toh ((int64_t) h->hdr->last_modified);
if (h->msglvl >= 2) {
char *name = _hivex_recode (h, utf16le_to_utf8,
h->hdr->name, 64, NULL);
fprintf (stderr,
"hivex_open: header fields:\n"
" file version %" PRIu32 ".%" PRIu32 "\n"
" sequence nos %" PRIu32 " %" PRIu32 "\n"
" (sequences nos should match if hive was synched at shutdown)\n"
" last modified %" PRIi64 "\n"
" (Windows filetime, x 100 ns since 1601-01-01)\n"
" original file name %s\n"
" (only 32 chars are stored, name is probably truncated)\n"
" root offset 0x%x + 0x1000\n"
" end of last page 0x%x + 0x1000 (total file size 0x%zx)\n"
" checksum 0x%x (calculated 0x%x)\n",
major_ver, le32toh (h->hdr->minor_ver),
le32toh (h->hdr->sequence1), le32toh (h->hdr->sequence2),
h->last_modified,
name ? name : "(conversion failed)",
le32toh (h->hdr->offset),
le32toh (h->hdr->blocks), h->size,
le32toh (h->hdr->csum), sum);
free (name);
}
h->rootoffs = le32toh (h->hdr->offset) + 0x1000;
h->endpages = le32toh (h->hdr->blocks) + 0x1000;
DEBUG (2, "root offset = 0x%zx", h->rootoffs);
/* We'll set this flag when we see a block with the root offset (ie.
* the root block).
*/
int seen_root_block = 0, bad_root_block = 0;
/* Collect some stats. */
size_t pages = 0; /* Number of hbin pages read. */
size_t smallest_page = SIZE_MAX, largest_page = 0;
size_t blocks = 0; /* Total number of blocks found. */
size_t smallest_block = SIZE_MAX, largest_block = 0, blocks_bytes = 0;
size_t used_blocks = 0; /* Total number of used blocks found. */
size_t used_size = 0; /* Total size (bytes) of used blocks. */
/* Read the pages and blocks. The aim here is to be robust against
* corrupt or malicious registries. So we make sure the loops
* always make forward progress. We add the address of each block
* we read to a hash table so pointers will only reference the start
* of valid blocks.
*/
size_t off;
struct ntreg_hbin_page *page;
for (off = 0x1000; off < h->size; off += le32toh (page->page_size)) {
if (off >= h->endpages)
break;
page = (struct ntreg_hbin_page *) ((char *) h->addr + off);
if (page->magic[0] != 'h' ||
page->magic[1] != 'b' ||
page->magic[2] != 'i' ||
page->magic[3] != 'n') {
if (!h->unsafe) {
SET_ERRNO (ENOTSUP,
"%s: trailing garbage at end of file "
"(at 0x%zx, after %zu pages)",
filename, off, pages);
goto error;
}
DEBUG (2,
"page not found at expected offset 0x%zx, "
"seeking until one is found or EOF is reached",
off);
int found = 0;
while (off < h->size) {
off += 0x1000;
if (off >= h->endpages)
break;
page = (struct ntreg_hbin_page *) ((char *) h->addr + off);
if (page->magic[0] == 'h' &&
page->magic[1] == 'b' &&
page->magic[2] == 'i' &&
page->magic[3] == 'n') {
DEBUG (2, "found next page by seeking at 0x%zx", off);
found = 1;
break;
}
}
if (!found) {
DEBUG (2, "page not found and end of pages section reached");
break;
}
}
size_t page_size = le32toh (page->page_size);
DEBUG (2, "page at 0x%zx, size %zu", off, page_size);
pages++;
if (page_size < smallest_page) smallest_page = page_size;
if (page_size > largest_page) largest_page = page_size;
if (page_size <= sizeof (struct ntreg_hbin_page) ||
(page_size & 0x0fff) != 0) {
SET_ERRNO (ENOTSUP,
"%s: page size %zu at 0x%zx, bad registry",
filename, page_size, off);
goto error;
}
if (off + page_size > h->size) {
SET_ERRNO (ENOTSUP,
"%s: page size %zu at 0x%zx extends beyond end of file, bad registry",
filename, page_size, off);
goto error;
}
size_t page_offset = le32toh(page->offset_first) + 0x1000;
if (page_offset != off) {
SET_ERRNO (ENOTSUP,
"%s: declared page offset (0x%zx) does not match computed "
"offset (0x%zx), bad registry",
filename, page_offset, off);
goto error;
}
/* Read the blocks in this page. */
size_t blkoff;
struct ntreg_hbin_block *block;
size_t seg_len;
for (blkoff = off + 0x20;
blkoff < off + page_size;
blkoff += seg_len) {
blocks++;
int is_root = blkoff == h->rootoffs;
if (is_root)
seen_root_block = 1;
block = (struct ntreg_hbin_block *) ((char *) h->addr + blkoff);
int used;
seg_len = block_len (h, blkoff, &used);
/* https://gcc.gnu.org/bugzilla/show_bug.cgi?id=78665 */
#pragma GCC diagnostic push
#pragma GCC diagnostic ignored "-Wstrict-overflow"
if (seg_len <= 4 || (seg_len & 3) != 0) {
#pragma GCC diagnostic pop
if (is_root || !h->unsafe) {
SET_ERRNO (ENOTSUP,
"%s, the block at 0x%zx has invalid size %" PRIu32
", bad registry",
filename, blkoff, le32toh (block->seg_len));
goto error;
} else {
DEBUG (2,
"%s: block at 0x%zx has invalid size %" PRIu32 ", skipping",
filename, blkoff, le32toh (block->seg_len));
break;
}
}
if (h->msglvl >= 2) {
unsigned char *id = (unsigned char *) block->id;
int id0 = id[0], id1 = id[1];
fprintf (stderr, "%s: %s: "
"%s block id %d,%d (%c%c) at 0x%zx size %zu%s\n",
"hivex", __func__,
used ? "used" : "free",
id0, id1,
c_isprint (id0) ? id0 : '.',
c_isprint (id1) ? id1 : '.',
blkoff,
seg_len, is_root ? " (root)" : "");
}
blocks_bytes += seg_len;
if (seg_len < smallest_block) smallest_block = seg_len;
if (seg_len > largest_block) largest_block = seg_len;
if (is_root && !used)
bad_root_block = 1;
if (used) {
used_blocks++;
used_size += seg_len;
/* Root block must be an nk-block. */
if (is_root && (block->id[0] != 'n' || block->id[1] != 'k'))
bad_root_block = 1;
/* Note this blkoff is a valid address. */
BITMAP_SET (h->bitmap, blkoff);
}
}
}
if (!seen_root_block) {
SET_ERRNO (ENOTSUP, "%s: no root block found", filename);
goto error;
}
if (bad_root_block) {
SET_ERRNO (ENOTSUP, "%s: bad root block (free or not nk)", filename);
goto error;
}
DEBUG (1, "successfully read Windows Registry hive file:\n"
" pages: %zu [sml: %zu, lge: %zu]\n"
" blocks: %zu [sml: %zu, avg: %zu, lge: %zu]\n"
" blocks used: %zu\n"
" bytes used: %zu",
pages, smallest_page, largest_page,
blocks, smallest_block, blocks_bytes / blocks, largest_block,
used_blocks, used_size);
return h;
error:;
int err = errno;
if (h) {
free (h->bitmap);
if (h->addr && h->size && h->addr != MAP_FAILED) {
if (!h->writable)
munmap (h->addr, h->size);
else
free (h->addr);
}
if (h->fd >= 0)
close (h->fd);
free (h->filename);
free (h);
}
errno = err;
return NULL;
}
/*-------------------------------------------------------------------------
*
* Preform initialization of guests and guest CPUs
*
* Should be called on BSP only while all APs are stopped
*
* Return TRUE for success
*
*------------------------------------------------------------------------- */
boolean_t initialize_all_guests(uint32_t number_of_host_processors,
const mon_memory_layout_t *mon_memory_layout,
const mon_guest_startup_t *
primary_guest_startup_state,
uint32_t number_of_secondary_guests,
const mon_guest_startup_t *
secondary_guests_startup_state_array,
const mon_application_params_struct_t *
application_params)
{
guest_handle_t primary_guest;
gpm_handle_t primary_guest_startup_gpm;
boolean_t ok = FALSE;
/* guest_handle_t cur_guest; */
guest_cpu_handle_t gcpu;
guest_gcpu_econtext_t gcpu_context;
MON_ASSERT(hw_cpu_id() == 0);
MON_ASSERT(number_of_host_processors > 0);
MON_ASSERT(mon_memory_layout);
MON_ASSERT(primary_guest_startup_state);
if (number_of_secondary_guests > 0) {
MON_LOG(mask_anonymous, level_trace,
"initialize_all_guests ASSERT: Secondary guests are"
" yet not implemented\n");
MON_ASSERT(secondary_guests_startup_state_array);
/* init guests and allocate memory for them */
/* shutdown temporary layout object */
MON_DEADLOOP();
return FALSE;
}
/* first init primary guest */
MON_LOG(mask_anonymous, level_trace, "Init primary guest\n");
/* BUGBUG: This is a workaround until loader will not do this!!! */
BITMAP_SET(((mon_guest_startup_t *)primary_guest_startup_state)->flags,
MON_GUEST_FLAG_REAL_BIOS_ACCESS_ENABLE |
MON_GUEST_FLAG_LAUNCH_IMMEDIATELY);
/* TODO: Uses global policym but should be part of mon_guest_startup_t
* structure. */
primary_guest = init_single_guest(number_of_host_processors,
primary_guest_startup_state, NULL);
if (!primary_guest) {
MON_LOG(mask_anonymous, level_trace,
"initialize_all_guests: Cannot init primary guest\n");
MON_DEADLOOP();
return FALSE;
}
guest_set_primary(primary_guest);
primary_guest_startup_gpm = mon_guest_get_startup_gpm(primary_guest);
/* init memory layout in the startup gpm */
ok = init_memory_layout(mon_memory_layout,
primary_guest_startup_gpm,
number_of_secondary_guests > 0,
application_params);
/* Set active_gpm to startup gpm */
for (gcpu = mon_guest_gcpu_first(primary_guest, &gcpu_context); gcpu;
gcpu = mon_guest_gcpu_next(&gcpu_context))
mon_gcpu_set_current_gpm(gcpu, primary_guest_startup_gpm);
MON_LOG(mask_anonymous, level_trace,
"Primary guest initialized successfully\n");
return TRUE;
}