zte's code,first commit

Change-Id: I9a04da59e459a9bc0d67f101f700d9d7dc8d681b
diff --git a/ap/app/busybox/src/util-linux/volume_id/util.c b/ap/app/busybox/src/util-linux/volume_id/util.c
new file mode 100644
index 0000000..061545f
--- /dev/null
+++ b/ap/app/busybox/src/util-linux/volume_id/util.c
@@ -0,0 +1,265 @@
+/*
+ * volume_id - reads filesystem label and uuid
+ *
+ * Copyright (C) 2005 Kay Sievers <kay.sievers@vrfy.org>
+ *
+ *	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, write to the Free Software
+ *	Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
+ */
+
+#include "volume_id_internal.h"
+
+void volume_id_set_unicode16(char *str, size_t len, const uint8_t *buf, enum endian endianess, size_t count)
+{
+	unsigned i, j;
+	unsigned c;
+
+	j = 0;
+	for (i = 0; i + 2 <= count; i += 2) {
+		if (endianess == LE)
+			c = (buf[i+1] << 8) | buf[i];
+		else
+			c = (buf[i] << 8) | buf[i+1];
+		if (c == 0)
+			break;
+		if (j+1 >= len)
+			break;
+		if (c < 0x80) {
+			/* 0xxxxxxx */
+		} else {
+			uint8_t topbits = 0xc0;
+			if (j+2 >= len)
+				break;
+			if (c < 0x800) {
+				/* 110yyyxx 10xxxxxx */
+			} else {
+				if (j+3 >= len)
+					break;
+				/* 1110yyyy 10yyyyxx 10xxxxxx */
+				str[j++] = (uint8_t) (0xe0 | (c >> 12));
+				topbits = 0x80;
+			}
+			str[j++] = (uint8_t) (topbits | ((c >> 6) & 0x3f));
+			c = 0x80 | (c & 0x3f);
+		}
+		str[j++] = (uint8_t) c;
+	}
+	str[j] = '\0';
+}
+
+#ifdef UNUSED
+static const char *usage_to_string(enum volume_id_usage usage_id)
+{
+	switch (usage_id) {
+	case VOLUME_ID_FILESYSTEM:
+		return "filesystem";
+	case VOLUME_ID_PARTITIONTABLE:
+		return "partitiontable";
+	case VOLUME_ID_OTHER:
+		return "other";
+	case VOLUME_ID_RAID:
+		return "raid";
+	case VOLUME_ID_DISKLABEL:
+		return "disklabel";
+	case VOLUME_ID_CRYPTO:
+		return "crypto";
+	case VOLUME_ID_UNPROBED:
+		return "unprobed";
+	case VOLUME_ID_UNUSED:
+		return "unused";
+	}
+	return NULL;
+}
+
+void volume_id_set_usage_part(struct volume_id_partition *part, enum volume_id_usage usage_id)
+{
+	part->usage_id = usage_id;
+	part->usage = usage_to_string(usage_id);
+}
+
+void volume_id_set_usage(struct volume_id *id, enum volume_id_usage usage_id)
+{
+	id->usage_id = usage_id;
+	id->usage = usage_to_string(usage_id);
+}
+
+void volume_id_set_label_raw(struct volume_id *id, const uint8_t *buf, size_t count)
+{
+	memcpy(id->label_raw, buf, count);
+	id->label_raw_len = count;
+}
+#endif
+
+#ifdef NOT_NEEDED
+static size_t strnlen(const char *s, size_t maxlen)
+{
+	size_t i;
+	if (!maxlen) return 0;
+	if (!s) return 0;
+	for (i = 0; *s && i < maxlen; ++s) ++i;
+	return i;
+}
+#endif
+
+void volume_id_set_label_string(struct volume_id *id, const uint8_t *buf, size_t count)
+{
+	unsigned i;
+
+	memcpy(id->label, buf, count);
+
+	/* remove trailing whitespace */
+	i = strnlen(id->label, count);
+	while (i--) {
+		if (!isspace(id->label[i]))
+			break;
+	}
+	id->label[i+1] = '\0';
+}
+
+void volume_id_set_label_unicode16(struct volume_id *id, const uint8_t *buf, enum endian endianess, size_t count)
+{
+	volume_id_set_unicode16(id->label, sizeof(id->label), buf, endianess, count);
+}
+
+void volume_id_set_uuid(struct volume_id *id, const uint8_t *buf, enum uuid_format format)
+{
+	unsigned i;
+	unsigned count = (format == UUID_DCE_STRING ? VOLUME_ID_UUID_SIZE : 4 << format);
+
+//	memcpy(id->uuid_raw, buf, count);
+//	id->uuid_raw_len = count;
+
+	/* if set, create string in the same format, the native platform uses */
+	for (i = 0; i < count; i++)
+		if (buf[i] != 0)
+			goto set;
+	return; /* all bytes are zero, leave it empty ("") */
+
+set:
+	switch (format) {
+	case UUID_DOS:
+		sprintf(id->uuid, "%02X%02X-%02X%02X",
+			buf[3], buf[2], buf[1], buf[0]);
+		break;
+	case UUID_NTFS:
+		sprintf(id->uuid, "%02X%02X%02X%02X%02X%02X%02X%02X",
+			buf[7], buf[6], buf[5], buf[4],
+			buf[3], buf[2], buf[1], buf[0]);
+		break;
+	case UUID_DCE:
+		sprintf(id->uuid,
+			"%02x%02x%02x%02x-%02x%02x-%02x%02x-%02x%02x-%02x%02x%02x%02x%02x%02x",
+			buf[0], buf[1], buf[2], buf[3],
+			buf[4], buf[5],
+			buf[6], buf[7],
+			buf[8], buf[9],
+			buf[10], buf[11], buf[12], buf[13], buf[14], buf[15]);
+		break;
+	case UUID_DCE_STRING:
+		memcpy(id->uuid, buf, count);
+		id->uuid[count] = '\0';
+		break;
+	}
+}
+
+/* Do not use xlseek here. With it, single corrupted filesystem
+ * may result in attempt to seek past device -> exit.
+ * It's better to ignore such fs and continue.  */
+void *volume_id_get_buffer(struct volume_id *id, uint64_t off, size_t len)
+{
+	uint8_t *dst;
+	unsigned small_off;
+	ssize_t read_len;
+
+	dbg("get buffer off 0x%llx(%llu), len 0x%zx",
+		(unsigned long long) off, (unsigned long long) off, len);
+
+	/* check if requested area fits in superblock buffer */
+	if (off + len <= SB_BUFFER_SIZE
+	 /* && off <= SB_BUFFER_SIZE - want this paranoid overflow check? */
+	) {
+		if (id->sbbuf == NULL) {
+			id->sbbuf = xmalloc(SB_BUFFER_SIZE);
+		}
+		small_off = off;
+		dst = id->sbbuf;
+
+		/* check if we need to read */
+		len += off;
+		if (len <= id->sbbuf_len)
+			goto ret; /* we already have it */
+
+		dbg("read sbbuf len:0x%x", (unsigned) len);
+		id->sbbuf_len = len;
+		off = 0;
+		goto do_read;
+	}
+
+	if (len > SEEK_BUFFER_SIZE) {
+		dbg("seek buffer too small %d", SEEK_BUFFER_SIZE);
+		return NULL;
+	}
+	dst = id->seekbuf;
+
+	/* check if we need to read */
+	if ((off >= id->seekbuf_off)
+	 && ((off + len) <= (id->seekbuf_off + id->seekbuf_len))
+	) {
+		small_off = off - id->seekbuf_off; /* can't overflow */
+		goto ret; /* we already have it */
+	}
+
+	id->seekbuf_off = off;
+	id->seekbuf_len = len;
+	id->seekbuf = xrealloc(id->seekbuf, len);
+	small_off = 0;
+	dst = id->seekbuf;
+	dbg("read seekbuf off:0x%llx len:0x%zx",
+				(unsigned long long) off, len);
+ do_read:
+	if (lseek(id->fd, off, SEEK_SET) != off) {
+		dbg("seek(0x%llx) failed", (unsigned long long) off);
+		goto err;
+	}
+	read_len = full_read(id->fd, dst, len);
+	if (read_len != len) {
+		dbg("requested 0x%x bytes, got 0x%x bytes",
+				(unsigned) len, (unsigned) read_len);
+ err:
+		/* No filesystem can be this tiny. It's most likely
+		 * non-associated loop device, empty drive and so on.
+		 * Flag it, making it possible to short circuit future
+		 * accesses. Rationale:
+		 * users complained of slow blkid due to empty floppy drives.
+		 */
+		if (off < 64*1024)
+			id->error = 1;
+		/* id->seekbuf_len or id->sbbuf_len is wrong now! Fixing. */
+		volume_id_free_buffer(id);
+		return NULL;
+	}
+ ret:
+	return dst + small_off;
+}
+
+void volume_id_free_buffer(struct volume_id *id)
+{
+	free(id->sbbuf);
+	id->sbbuf = NULL;
+	id->sbbuf_len = 0;
+	free(id->seekbuf);
+	id->seekbuf = NULL;
+	id->seekbuf_len = 0;
+	id->seekbuf_off = 0; /* paranoia */
+}