o Replace unicode16() by utf8_to_utf16().

o  Introduce utf16_to_utf8().
o  Add option -l to the show command to display the GPT label instead
   of the friendly partition type.
o  Add option -u to the show command to suppress the friendly output
   and print th raw UUIDs instead.

4 files changed, 139 insertions, 19 deletions

diff --git a/sbin/gpt/gpt.c b/sbin/gpt/gpt.c
index 3a114c1..29a1983 100644
--- a/sbin/gpt/gpt.c
+++ b/sbin/gpt/gpt.c
@@ -118,13 +118,112 @@ crc32(const void *buf, size_t size)
 	return crc ^ ~0U;
 }
 
+uint8_t *
+utf16_to_utf8(uint16_t *s16)
+{
+	static uint8_t *s8 = NULL;
+	static size_t s8len = 0;
+	size_t s8idx, s16idx, s16len;
+	uint32_t utfchar;
+	unsigned int c;
+
+	s16len = 0;
+	while (s16[s16len++] != 0)
+		;
+	if (s8len < s16len * 3) {
+		if (s8 != NULL)
+			free(s8);
+		s8len = s16len * 3;
+		s8 = calloc(s16len, 3);
+	}
+	s8idx = s16idx = 0;
+	while (s16idx < s16len) {
+		utfchar = le16toh(s16[s16idx++]);
+		if ((utfchar & 0xf800) == 0xd800) {
+			c = le16toh(s16[s16idx]);
+			if ((utfchar & 0x400) != 0 || (c & 0xfc00) != 0xdc00)
+				utfchar = 0xfffd;
+			else
+				s16idx++;
+		}
+		if (utfchar < 0x80) {
+			s8[s8idx++] = utfchar;
+		} else if (utfchar < 0x800) {
+			s8[s8idx++] = 0xc0 | (utfchar >> 6);
+			s8[s8idx++] = 0x80 | (utfchar & 0x3f);
+		} else if (utfchar < 0x10000) {
+			s8[s8idx++] = 0xe0 | (utfchar >> 12);
+			s8[s8idx++] = 0x80 | ((utfchar >> 6) & 0x3f);
+			s8[s8idx++] = 0x80 | (utfchar & 0x3f);
+		} else if (utfchar < 0x200000) {
+			s8[s8idx++] = 0xf0 | (utfchar >> 18);
+			s8[s8idx++] = 0x80 | ((utfchar >> 12) & 0x3f);
+			s8[s8idx++] = 0x80 | ((utfchar >> 6) & 0x3f);
+			s8[s8idx++] = 0x80 | (utfchar & 0x3f);
+		}
+	}
+	return (s8);
+}
+
 void
-unicode16(short *dst, const wchar_t *src, size_t len)
+utf8_to_utf16(const uint8_t *s8, uint16_t *s16, size_t s16len)
 {
-	while (len-- && *src != 0)
-		*dst++ = *src++;
-	if (len)
-		*dst = 0;
+	size_t s16idx, s8idx, s8len;
+	uint32_t utfchar;
+	unsigned int c, utfbytes;
+
+	s8len = 0;
+	while (s8[s8len++] != 0)
+		;
+	s8idx = s16idx = 0;
+	utfbytes = 0;
+	do {
+		c = s8[s8idx++];
+		if ((c & 0xc0) != 0x80) {
+			/* Initial characters. */
+			if (utfbytes != 0) {
+				/* Incomplete encoding. */
+				s16[s16idx++] = 0xfffd;
+				if (s16idx == s16len) {
+					s16[--s16idx] = 0;
+					return;
+				}
+			}
+			if ((c & 0xf8) == 0xf0) {
+				utfchar = c & 0x07;
+				utfbytes = 3;
+			} else if ((c & 0xf0) == 0xe0) {
+				utfchar = c & 0x0f;
+				utfbytes = 2;
+			} else if ((c & 0xe0) == 0xc0) {
+				utfchar = c & 0x1f;
+				utfbytes = 1;
+			} else {
+				utfchar = c & 0x7f;
+				utfbytes = 0;
+			}
+		} else {
+			/* Followup characters. */
+			if (utfbytes > 0) {
+				utfchar = (utfchar << 6) + (c & 0x3f);
+				utfbytes--;
+			} else if (utfbytes == 0)
+				utfbytes = -1;
+		}
+		if (utfbytes == 0) {
+			if (utfchar >= 0x10000 && s16idx + 2 >= s16len)
+				utfchar = 0xfffd;
+			if (utfchar >= 0x10000) {
+				s16[s16idx++] = 0xd800 | ((utfchar>>10)-0x40);
+				s16[s16idx++] = 0xdc00 | (utfchar & 0x3ff);
+			} else
+				s16[s16idx++] = utfchar;
+			if (s16idx == s16len) {
+				s16[--s16idx] = 0;
+				return;
+			}
+		}
+	} while (c != 0);
 }
 
 void
diff --git a/sbin/gpt/gpt.h b/sbin/gpt/gpt.h
index baa651f..596a68c 100644
--- a/sbin/gpt/gpt.h
+++ b/sbin/gpt/gpt.h
@@ -70,7 +70,9 @@ void	gpt_close(int);
 int	gpt_open(const char *);
 void*	gpt_read(int, off_t, size_t);
 int	gpt_write(int, map_t *);
-void	unicode16(short *, const wchar_t *, size_t);
+
+uint8_t *utf16_to_utf8(uint16_t *);
+void	utf8_to_utf16(const uint8_t *, uint16_t *, size_t);
 
 int	cmd_add(int, char *[]);
 int	cmd_create(int, char *[]);
diff --git a/sbin/gpt/migrate.c b/sbin/gpt/migrate.c
index 5322792..fb36d7a 100644
--- a/sbin/gpt/migrate.c
+++ b/sbin/gpt/migrate.c
@@ -100,22 +100,22 @@ migrate_disklabel(int fd, off_t start, struct gpt_ent *ent)
 		case FS_SWAP: {
 			uuid_t swap = GPT_ENT_TYPE_FREEBSD_SWAP;
 			le_uuid_enc(&ent->ent_type, &swap);
-			unicode16(ent->ent_name,
-			    L"FreeBSD swap partition", 36);
+			utf8_to_utf16("FreeBSD swap partition",
+			    ent->ent_name, 36);
 			break;
 		}
 		case FS_BSDFFS: {
 			uuid_t ufs = GPT_ENT_TYPE_FREEBSD_UFS;
 			le_uuid_enc(&ent->ent_type, &ufs);
-			unicode16(ent->ent_name,
-			    L"FreeBSD UFS partition", 36);
+			utf8_to_utf16("FreeBSD UFS partition",
+			    ent->ent_name, 36);
 			break;
 		}
 		case FS_VINUM: {
 			uuid_t vinum = GPT_ENT_TYPE_FREEBSD_VINUM;
 			le_uuid_enc(&ent->ent_type, &vinum);
-			unicode16(ent->ent_name,
-			    L"FreeBSD vinum partition", 36);
+			utf8_to_utf16("FreeBSD vinum partition",
+			    ent->ent_name, 36);
 			break;
 		}
 		default:
@@ -255,8 +255,8 @@ migrate(int fd)
 				le_uuid_enc(&ent->ent_type, &freebsd);
 				ent->ent_lba_start = htole64((uint64_t)start);
 				ent->ent_lba_end = htole64(start + size - 1LL);
-				unicode16(ent->ent_name,
-				    L"FreeBSD disklabel partition", 36);
+				utf8_to_utf16("FreeBSD disklabel partition",
+				    ent->ent_name, 36);
 				ent++;
 			} else
 				ent = migrate_disklabel(fd, start, ent);
@@ -267,7 +267,8 @@ migrate(int fd)
 			le_uuid_enc(&ent->ent_type, &efi_slice);
 			ent->ent_lba_start = htole64((uint64_t)start);
 			ent->ent_lba_end = htole64(start + size - 1LL);
-			unicode16(ent->ent_name, L"EFI system partition", 36);
+			utf8_to_utf16("EFI system partition",
+			    ent->ent_name, 36);
 			ent++;
 			break;
 		}
diff --git a/sbin/gpt/show.c b/sbin/gpt/show.c
index f2c94e1..4df3be1 100644
--- a/sbin/gpt/show.c
+++ b/sbin/gpt/show.c
@@ -39,12 +39,15 @@ __FBSDID("$FreeBSD$");
 #include "map.h"
 #include "gpt.h"
 
+static int show_label = 0;
+static int show_uuid = 0;
+
 static void
 usage_show(void)
 {
 
 	fprintf(stderr,
-	    "usage: %s device ...\n", getprogname());
+	    "usage: %s [-lu] device ...\n", getprogname());
 	exit(1);
 }
 
@@ -62,6 +65,9 @@ friendly(uuid_t *t)
 	static char buf[80];
 	char *s;
 
+	if (show_uuid)
+		goto unfriendly;
+
 	if (uuid_equal(t, &efi_slice, NULL))
 		return ("EFI System");
 	if (uuid_equal(t, &swap, NULL))
@@ -80,6 +86,7 @@ friendly(uuid_t *t)
 	if (uuid_equal(t, &msr, NULL))
 		return ("Windows reserved");
 
+unfriendly:
 	uuid_to_string(t, &s, NULL);
 	strlcpy(buf, s, sizeof buf);
 	free(s);
@@ -148,8 +155,13 @@ show(int fd __unused)
 		case MAP_TYPE_GPT_PART:
 			printf("GPT part ");
 			ent = m->map_data;
-			le_uuid_dec(&ent->ent_type, &type);
-			printf("- %s", friendly(&type));
+			if (show_label) {
+				printf("- \"%s\"",
+				    utf16_to_utf8(ent->ent_name));
+			} else {
+				le_uuid_dec(&ent->ent_type, &type);
+				printf("- %s", friendly(&type));
+			}
 			break;
 		case MAP_TYPE_PMBR:
 			printf("PMBR");
@@ -165,8 +177,14 @@ cmd_show(int argc, char *argv[])
 {
 	int ch, fd;
 
-	while ((ch = getopt(argc, argv, "")) != -1) {
+	while ((ch = getopt(argc, argv, "lu")) != -1) {
 		switch(ch) {
+		case 'l':
+			show_label = 1;
+			break;
+		case 'u':
+			show_uuid = 1;
+			break;
 		default:
 			usage_show();
 		}