speakup: Add /dev/synthu device

/dev/synth has always been 8bit, but applications nowadays mostly
expect to be using utf-8 encoding.  This adds /dev/synthu to be able
to synthesize non-latin1 characters.  This however remains limited
to 16bit unicode like the rest of speakup.  Any odd input or input
beyond 16bit is just discarded.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>

===================================================================
Link: https://lore.kernel.org/r/20240204155825.ditstifsbqndnce3@beginSigned-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

drivers/accessibility/speakup/devsynth.c

@@ -7,9 +7,10 @@
 #include "speakup.h"
 #include "spk_priv.h"
 
-static int misc_registered;
+static int synth_registered, synthu_registered;
 static int dev_opened;
 
+/* Latin1 version */
 static ssize_t speakup_file_write(struct file *fp, const char __user *buffer,
 				  size_t nbytes, loff_t *ppos)
 {
@@ -34,6 +35,97 @@ static ssize_t speakup_file_write(struct file *fp, const char __user *buffer,
 	return (ssize_t)nbytes;
 }
 
+/* UTF-8 version */
+static ssize_t speakup_file_writeu(struct file *fp, const char __user *buffer,
+				   size_t nbytes, loff_t *ppos)
+{
+	size_t count = nbytes, want;
+	const char __user *ptr = buffer;
+	size_t bytes;
+	unsigned long flags;
+	unsigned char buf[256];
+	u16 ubuf[256];
+	size_t in, in2, out;
+
+	if (!synth)
+		return -ENODEV;
+
+	want = 1;
+	while (count >= want) {
+		/* Copy some UTF-8 piece from userland */
+		bytes = min(count, sizeof(buf));
+		if (copy_from_user(buf, ptr, bytes))
+			return -EFAULT;
+
+		/* Convert to u16 */
+		for (in = 0, out = 0; in < bytes; in++) {
+			unsigned char c = buf[in];
+			int nbytes = 8 - fls(c ^ 0xff);
+			u32 value;
+
+			switch (nbytes) {
+			case 8: /* 0xff */
+			case 7: /* 0xfe */
+			case 1: /* 0x80 */
+				/* Invalid, drop */
+				goto drop;
+
+			case 0:
+				/* ASCII, copy */
+				ubuf[out++] = c;
+				continue;
+
+			default:
+				/* 2..6-byte UTF-8 */
+
+				if (bytes - in < nbytes) {
+					/* We don't have it all yet, stop here
+					 * and wait for the rest
+					 */
+					bytes = in;
+					want = nbytes;
+					continue;
+				}
+
+				/* First byte */
+				value = c & ((1u << (7 - nbytes)) - 1);
+
+				/* Other bytes */
+				for (in2 = 2; in2 <= nbytes; in2++) {
+					c = buf[in + 1];
+					if ((c & 0xc0) != 0x80)	{
+						/* Invalid, drop the head */
+						want = 1;
+						goto drop;
+					}
+					value = (value << 6) | (c & 0x3f);
+					in++;
+				}
+
+				if (value < 0x10000)
+					ubuf[out++] = value;
+				want = 1;
+				break;
+			}
+drop:
+		}
+
+		count -= bytes;
+		ptr += bytes;
+
+		/* And speak this up */
+		if (out) {
+			spin_lock_irqsave(&speakup_info.spinlock, flags);
+			for (in = 0; in < out; in++)
+				synth_buffer_add(ubuf[in]);
+			synth_start();
+			spin_unlock_irqrestore(&speakup_info.spinlock, flags);
+		}
+	}
+
+	return (ssize_t)(nbytes - count);
+}
+
 static ssize_t speakup_file_read(struct file *fp, char __user *buf,
 				 size_t nbytes, loff_t *ppos)
 {
@@ -62,31 +154,57 @@ static const struct file_operations synth_fops = {
 	.release = speakup_file_release,
 };
 
+static const struct file_operations synthu_fops = {
+	.read = speakup_file_read,
+	.write = speakup_file_writeu,
+	.open = speakup_file_open,
+	.release = speakup_file_release,
+};
+
 static struct miscdevice synth_device = {
 	.minor = MISC_DYNAMIC_MINOR,
 	.name = "synth",
 	.fops = &synth_fops,
 };
 
+static struct miscdevice synthu_device = {
+	.minor = MISC_DYNAMIC_MINOR,
+	.name = "synthu",
+	.fops = &synthu_fops,
+};
+
 void speakup_register_devsynth(void)
 {
-	if (misc_registered != 0)
-		return;
-/* zero it so if register fails, deregister will not ref invalid ptrs */
+	if (!synth_registered) {
 		if (misc_register(&synth_device)) {
 			pr_warn("Couldn't initialize miscdevice /dev/synth.\n");
 		} else {
 			pr_info("initialized device: /dev/synth, node (MAJOR %d, MINOR %d)\n",
 				MISC_MAJOR, synth_device.minor);
-		misc_registered = 1;
+			synth_registered = 1;
+		}
+	}
+	if (!synthu_registered) {
+		if (misc_register(&synthu_device)) {
+			pr_warn("Couldn't initialize miscdevice /dev/synthu.\n");
+		} else {
+			pr_info("initialized device: /dev/synthu, node (MAJOR %d, MINOR %d)\n",
+				MISC_MAJOR, synthu_device.minor);
+			synthu_registered = 1;
+		}
 	}
 }
 
 void speakup_unregister_devsynth(void)
 {
-	if (!misc_registered)
-		return;
+	if (synth_registered) {
 		pr_info("speakup: unregistering synth device /dev/synth\n");
 		misc_deregister(&synth_device);
-	misc_registered = 0;
+		synth_registered = 0;
+	}
+	if (synthu_registered) {
+		pr_info("speakup: unregistering synth device /dev/synthu\n");
+		misc_deregister(&synthu_device);
+		synthu_registered = 0;
+	}
 }