This adds /dev/softsynthu, along /dev/softsynth, which emits output in
UTF-8 encoding, thus allowing to support 16bit characters.  Most of the
code is shared, only the read function has to behave differently in
latin1 and in unicode mode.  Since Linux only supports 16bit characters,
we can just hardcode the UTF-8 encoding.

Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Chris Brannon <chris@the-brannons.com>

Index: linux-4.10/drivers/staging/speakup/speakup_soft.c
===================================================================
--- linux-4.10.orig/drivers/staging/speakup/speakup_soft.c
+++ linux-4.10/drivers/staging/speakup/speakup_soft.c
@@ -29,6 +29,7 @@
 
 #define DRV_VERSION "2.6"
 #define SOFTSYNTH_MINOR 26 /* might as well give it one more than /dev/synth */
+#define SOFTSYNTHU_MINOR 27 /* might as well give it one more than /dev/synth */
 #define PROCSPEECH 0x0d
 #define CLEAR_SYNTH 0x18
 
@@ -37,7 +38,7 @@ static void softsynth_release(void);
 static int softsynth_is_alive(struct spk_synth *synth);
 static unsigned char get_index(void);
 
-static struct miscdevice synth_device;
+static struct miscdevice synth_device, synthu_device;
 static int init_pos;
 static int misc_registered;
 
@@ -199,13 +200,13 @@ static int softsynth_close(struct inode
 	return 0;
 }
 
-static ssize_t softsynth_read(struct file *fp, char __user *buf, size_t count,
-			      loff_t *pos)
+static ssize_t softsynthx_read(struct file *fp, char __user *buf, size_t count,
+			       loff_t *pos, int unicode)
 {
 	int chars_sent = 0;
 	char __user *cp;
 	char *init;
-	char ch;
+	u16 ch;
 	int empty;
 	unsigned long flags;
 	DEFINE_WAIT(wait);
@@ -213,7 +214,8 @@ static ssize_t softsynth_read(struct fil
 	spin_lock_irqsave(&speakup_info.spinlock, flags);
 	while (1) {
 		prepare_to_wait(&speakup_event, &wait, TASK_INTERRUPTIBLE);
-		synth_buffer_skip_nonlatin1();
+		if (!unicode)
+			synth_buffer_skip_nonlatin1();
 		if (!synth_buffer_empty() || speakup_info.flushing)
 			break;
 		spin_unlock_irqrestore(&speakup_info.spinlock, flags);
@@ -232,23 +234,57 @@ static ssize_t softsynth_read(struct fil
 
 	cp = buf;
 	init = get_initstring();
-	while (chars_sent < count) {
+
+	/* Keep 3 bytes available for a 16bit UTF-8-encoded character */
+	while (chars_sent <= count - 3) {
 		if (speakup_info.flushing) {
 			speakup_info.flushing = 0;
 			ch = '\x18';
-		} else if (synth_buffer_empty()) {
-			break;
 		} else if (init[init_pos]) {
 			ch = init[init_pos++];
 		} else {
+			if (!unicode)
+				synth_buffer_skip_nonlatin1();
+			if (synth_buffer_empty())
+				break;
 			ch = synth_buffer_getc();
 		}
 		spin_unlock_irqrestore(&speakup_info.spinlock, flags);
-		if (copy_to_user(cp, &ch, 1))
-			return -EFAULT;
+
+		if ((!unicode && ch < 0x100) || (unicode && ch < 0x80)) {
+			u_char c = ch;
+
+			if (copy_to_user(cp, &c, 1))
+				return -EFAULT;
+
+			chars_sent++;
+			cp++;
+		} else if (unicode && ch < 0x800) {
+			u_char s[2] = {
+				0xc0 | (ch >> 6),
+				0x80 | (ch & 0x3f)
+			};
+
+			if (copy_to_user(cp, s, sizeof(s)))
+				return -EFAULT;
+
+			chars_sent += sizeof(s);
+			cp += sizeof(s);
+		} else if (unicode) {
+			u_char s[3] = {
+				0xe0 | (ch >> 12),
+				0x80 | ((ch >> 6) & 0x3f),
+				0x80 | (ch & 0x3f)
+			};
+
+			if (copy_to_user(cp, s, sizeof(s)))
+				return -EFAULT;
+
+			chars_sent += sizeof(s);
+			cp += sizeof(s);
+		}
+
 		spin_lock_irqsave(&speakup_info.spinlock, flags);
-		chars_sent++;
-		cp++;
 	}
 	*pos += chars_sent;
 	empty = synth_buffer_empty();
@@ -260,6 +296,18 @@ static ssize_t softsynth_read(struct fil
 	return chars_sent;
 }
 
+static ssize_t softsynth_read(struct file *fp, char __user *buf, size_t count,
+			      loff_t *pos)
+{
+	return softsynthx_read(fp, buf, count, pos, 0);
+}
+
+static ssize_t softsynthu_read(struct file *fp, char __user *buf, size_t count,
+			       loff_t *pos)
+{
+	return softsynthx_read(fp, buf, count, pos, 1);
+}
+
 static int last_index;
 
 static ssize_t softsynth_write(struct file *fp, const char __user *buf,
@@ -309,6 +357,15 @@ static const struct file_operations soft
 	.release = softsynth_close,
 };
 
+static const struct file_operations softsynthu_fops = {
+	.owner = THIS_MODULE,
+	.poll = softsynth_poll,
+	.read = softsynthu_read,
+	.write = softsynth_write,
+	.open = softsynth_open,
+	.release = softsynth_close,
+};
+
 static int softsynth_probe(struct spk_synth *synth)
 {
 	if (misc_registered != 0)
@@ -322,16 +379,28 @@ static int softsynth_probe(struct spk_sy
 		return -ENODEV;
 	}
 
+	memset(&synthu_device, 0, sizeof(synthu_device));
+	synthu_device.minor = SOFTSYNTHU_MINOR;
+	synthu_device.name = "softsynthu";
+	synthu_device.fops = &softsynthu_fops;
+	if (misc_register(&synthu_device)) {
+		pr_warn("Couldn't initialize miscdevice /dev/softsynth.\n");
+		return -ENODEV;
+	}
+
 	misc_registered = 1;
 	pr_info("initialized device: /dev/softsynth, node (MAJOR 10, MINOR 26)\n");
+	pr_info("initialized device: /dev/softsynthu, node (MAJOR 10, MINOR 27)\n");
 	return 0;
 }
 
 static void softsynth_release(void)
 {
 	misc_deregister(&synth_device);
+	misc_deregister(&synthu_device);
 	misc_registered = 0;
 	pr_info("unregistered /dev/softsynth\n");
+	pr_info("unregistered /dev/softsynthu\n");
 }
 
 static int softsynth_is_alive(struct spk_synth *synth)