diff --git a/drivers/char/n_tty.c b/drivers/char/n_tty.c
index efbfe961..a9bc576 100644
--- a/drivers/char/n_tty.c
+++ b/drivers/char/n_tty.c
@@ -62,6 +62,17 @@
 #define TTY_THRESHOLD_THROTTLE		128 /* now based on remaining room */
 #define TTY_THRESHOLD_UNTHROTTLE 	128
 
+/*
+ * Special byte codes used in the echo buffer to represent operations
+ * or special handling of characters.  Bytes in the echo buffer that
+ * are not part of such special blocks are treated as normal character
+ * codes.
+ */
+#define ECHO_OP_START 0xff
+#define ECHO_OP_MOVE_BACK_COL 0x80
+#define ECHO_OP_SET_CANON_COL 0x81
+#define ECHO_OP_ERASE_TAB 0x82
+
 static inline unsigned char *alloc_buf(void)
 {
 	gfp_t prio = in_interrupt() ? GFP_ATOMIC : GFP_KERNEL;
@@ -169,6 +180,7 @@
  *
  *	Locking: tty_read_lock for read fields.
  */
+
 static void reset_buffer_flags(struct tty_struct *tty)
 {
 	unsigned long flags;
@@ -176,6 +188,11 @@
 	spin_lock_irqsave(&tty->read_lock, flags);
 	tty->read_head = tty->read_tail = tty->read_cnt = 0;
 	spin_unlock_irqrestore(&tty->read_lock, flags);
+
+	mutex_lock(&tty->echo_lock);
+	tty->echo_pos = tty->echo_cnt = tty->echo_overrun = 0;
+	mutex_unlock(&tty->echo_lock);
+
 	tty->canon_head = tty->canon_data = tty->erasing = 0;
 	memset(&tty->read_flags, 0, sizeof tty->read_flags);
 	n_tty_set_room(tty);
@@ -266,89 +283,116 @@
 }
 
 /**
- *	opost			-	output post processor
+ *	do_output_char			-	output one character
+ *	@c: character (or partial unicode symbol)
+ *	@tty: terminal device
+ *	@space: space available in tty driver write buffer
+ *
+ *	This is a helper function that handles one output character
+ *	(including special characters like TAB, CR, LF, etc.),
+ *	putting the results in the tty driver's write buffer.
+ *
+ *	Note that Linux currently ignores TABDLY, CRDLY, VTDLY, FFDLY
+ *	and NLDLY.  They simply aren't relevant in the world today.
+ *	If you ever need them, add them here.
+ *
+ *	Returns the number of bytes of buffer space used or -1 if
+ *	no space left.
+ *
+ *	Locking: should be called under the output_lock to protect
+ *		 the column state and space left in the buffer
+ */
+
+static int do_output_char(unsigned char c, struct tty_struct *tty, int space)
+{
+	int	spaces;
+
+	if (!space)
+		return -1;
+	
+	switch (c) {
+	case '\n':
+		if (O_ONLRET(tty))
+			tty->column = 0;
+		if (O_ONLCR(tty)) {
+			if (space < 2)
+				return -1;
+			tty->canon_column = tty->column = 0;
+			tty_put_char(tty, '\r');
+			tty_put_char(tty, c);
+			return 2;
+		}
+		tty->canon_column = tty->column;
+		break;
+	case '\r':
+		if (O_ONOCR(tty) && tty->column == 0)
+			return 0;
+		if (O_OCRNL(tty)) {
+			c = '\n';
+			if (O_ONLRET(tty))
+				tty->canon_column = tty->column = 0;
+			break;
+		}
+		tty->canon_column = tty->column = 0;
+		break;
+	case '\t':
+		spaces = 8 - (tty->column & 7);
+		if (O_TABDLY(tty) == XTABS) {
+			if (space < spaces)
+				return -1;
+			tty->column += spaces;
+			tty->ops->write(tty, "        ", spaces);
+			return spaces;
+		}
+		tty->column += spaces;
+		break;
+	case '\b':
+		if (tty->column > 0)
+			tty->column--;
+		break;
+	default:
+		if (O_OLCUC(tty))
+			c = toupper(c);
+		if (!iscntrl(c) && !is_continuation(c, tty))
+			tty->column++;
+		break;
+	}
+
+	tty_put_char(tty, c);
+	return 1;
+}
+
+/**
+ *	process_output			-	output post processor
  *	@c: character (or partial unicode symbol)
  *	@tty: terminal device
  *
  *	Perform OPOST processing.  Returns -1 when the output device is
- *	full and the character must be retried. Note that Linux currently
- *	ignores TABDLY, CRDLY, VTDLY, FFDLY and NLDLY. They simply aren't
- *	relevant in the world today. If you ever need them, add them here.
+ *	full and the character must be retried.
  *
- *	Called from both the receive and transmit sides and can be called
- *	re-entrantly. Relies on lock_kernel() for tty->column state.
+ *	Locking: output_lock to protect column state and space left
+ *		 (also, this is called from n_tty_write under the
+ *		  tty layer write lock)
  */
 
-static int opost(unsigned char c, struct tty_struct *tty)
+static int process_output(unsigned char c, struct tty_struct *tty)
 {
-	int	space, spaces;
+	int	space, retval;
+
+	mutex_lock(&tty->output_lock);
 
 	space = tty_write_room(tty);
-	if (!space)
-		return -1;
+	retval = do_output_char(c, tty, space);
 
-	lock_kernel();
-	if (O_OPOST(tty)) {
-		switch (c) {
-		case '\n':
-			if (O_ONLRET(tty))
-				tty->column = 0;
-			if (O_ONLCR(tty)) {
-				if (space < 2) {
-					unlock_kernel();
-					return -1;
-				}
-				tty_put_char(tty, '\r');
-				tty->column = 0;
-			}
-			tty->canon_column = tty->column;
-			break;
-		case '\r':
-			if (O_ONOCR(tty) && tty->column == 0) {
-				unlock_kernel();
-				return 0;
-			}
-			if (O_OCRNL(tty)) {
-				c = '\n';
-				if (O_ONLRET(tty))
-					tty->canon_column = tty->column = 0;
-				break;
-			}
-			tty->canon_column = tty->column = 0;
-			break;
-		case '\t':
-			spaces = 8 - (tty->column & 7);
-			if (O_TABDLY(tty) == XTABS) {
-				if (space < spaces) {
-					unlock_kernel();
-					return -1;
-				}
-				tty->column += spaces;
-				tty->ops->write(tty, "        ", spaces);
-				unlock_kernel();
-				return 0;
-			}
-			tty->column += spaces;
-			break;
-		case '\b':
-			if (tty->column > 0)
-				tty->column--;
-			break;
-		default:
-			if (O_OLCUC(tty))
-				c = toupper(c);
-			if (!iscntrl(c) && !is_continuation(c, tty))
-				tty->column++;
-			break;
-		}
-	}
-	tty_put_char(tty, c);
-	unlock_kernel();
-	return 0;
+	mutex_unlock(&tty->output_lock);
+	if (retval < 0)
+		return -1;
+	else
+		return 0;
 }
 
 /**
- *	opost_block		-	block postprocess
+ *	process_output_block		-	block post processor
  *	@tty: terminal device
  *	@inbuf: user buffer
  *	@nr: number of bytes
@@ -358,24 +402,29 @@
  *	the simple cases normally found and helps to generate blocks of
  *	symbols for the console driver and thus improve performance.
  *
- *	Called from n_tty_write under the tty layer write lock. Relies
- *	on lock_kernel for the tty->column state.
+ *	Locking: output_lock to protect column state and space left
+ *		 (also, this is called from n_tty_write under the
+ *		  tty layer write lock)
  */
 
-static ssize_t opost_block(struct tty_struct *tty,
-		       const unsigned char *buf, unsigned int nr)
+static ssize_t process_output_block(struct tty_struct *tty,
+				    const unsigned char *buf, unsigned int nr)
 {
 	int	space;
 	int 	i;
 	const unsigned char *cp;
 
+	mutex_lock(&tty->output_lock);
+
 	space = tty_write_room(tty);
 	if (!space)
+	{
+		mutex_unlock(&tty->output_lock);
 		return 0;
+	}
 	if (nr > space)
 		nr = space;
 
-	lock_kernel();
 	for (i = 0, cp = buf; i < nr; i++, cp++) {
 		switch (*cp) {
 		case '\n':
@@ -407,46 +456,393 @@
 		}
 	}
 break_out:
-	if (tty->ops->flush_chars)
-		tty->ops->flush_chars(tty);
 	i = tty->ops->write(tty, buf, i);
-	unlock_kernel();
+
+	mutex_unlock(&tty->output_lock);
 	return i;
 }
 
+/**
+ *	process_echoes	-	write pending echo characters
+ *	@tty: terminal device
+ *
+ *	Write previously buffered echo (and other ldisc-generated)
+ *	characters to the tty.
+ *
+ *	Characters generated by the ldisc (including echoes) need to
+ *	be buffered because the driver's write buffer can fill during
+ *	heavy program output.  Echoing straight to the driver will
+ *	often fail under these conditions, causing lost characters and
+ *	resulting mismatches of ldisc state information.
+ *
+ *	Since the ldisc state must represent the characters actually sent
+ *	to the driver at the time of the write, operations like certain
+ *	changes in column state are also saved in the buffer and executed
+ *	here.
+ *
+ *	A circular fifo buffer is used so that the most recent characters
+ *	are prioritized.  Also, when control characters are echoed with a
+ *	prefixed "^", the pair is treated atomically and thus not separated.
+ *
+ *	Locking: output_lock to protect column state and space left,
+ *		 echo_lock to protect the echo buffer
+ */
+
+static void process_echoes(struct tty_struct *tty)
+{
+	int	space, nr;
+	unsigned char c;
+	unsigned char *cp, *buf_end;
+
+	if (!tty->echo_cnt)
+		return;
+
+	mutex_lock(&tty->output_lock);
+	mutex_lock(&tty->echo_lock);
+
+	space = tty_write_room(tty);
+
+	buf_end = tty->echo_buf + N_TTY_BUF_SIZE;
+	cp = tty->echo_buf + tty->echo_pos;
+	nr = tty->echo_cnt;
+	while (nr > 0) {
+		c = *cp;
+		if (c == ECHO_OP_START) {
+			unsigned char op;
+			unsigned char *opp;
+			int no_space_left = 0;
+
+			/*
+			 * If the buffer byte is the start of a multi-byte
+			 * operation, get the next byte, which is either the
+			 * op code or a control character value.
+			 */
+			opp = cp + 1;
+			if (opp == buf_end)
+				opp -= N_TTY_BUF_SIZE;
+			op = *opp;
+			
+			switch (op) {
+				unsigned int num_chars, num_bs;
+
+			case ECHO_OP_ERASE_TAB:
+				if (++opp == buf_end)
+					opp -= N_TTY_BUF_SIZE;
+				num_chars = *opp;
+
+				/*
+				 * Determine how many columns to go back
+				 * in order to erase the tab.
+				 * This depends on the number of columns
+				 * used by other characters within the tab
+				 * area.  If this (modulo 8) count is from
+				 * the start of input rather than from a
+				 * previous tab, we offset by canon column.
+				 * Otherwise, tab spacing is normal.
+				 */
+				if (!(num_chars & 0x80))
+					num_chars += tty->canon_column;
+				num_bs = 8 - (num_chars & 7);
+
+				if (num_bs > space) {
+					no_space_left = 1;
+					break;
+				}
+				space -= num_bs;
+				while (num_bs--) {
+					tty_put_char(tty, '\b');
+					if (tty->column > 0)
+						tty->column--;
+				}
+				cp += 3;
+				nr -= 3;
+				break;
+
+			case ECHO_OP_SET_CANON_COL:
+				tty->canon_column = tty->column;
+				cp += 2;
+				nr -= 2;
+				break;
+
+			case ECHO_OP_MOVE_BACK_COL:
+				if (tty->column > 0)
+					tty->column--;
+				cp += 2;
+				nr -= 2;
+				break;
+
+			case ECHO_OP_START:
+				/* This is an escaped echo op start code */
+				if (!space) {
+					no_space_left = 1;
+					break;
+				}
+				tty_put_char(tty, ECHO_OP_START);
+				tty->column++;
+				space--;
+				cp += 2;
+				nr -= 2;
+				break;
+
+			default:
+				if (iscntrl(op)) {
+					if (L_ECHOCTL(tty)) {
+						/*
+						 * Ensure there is enough space
+						 * for the whole ctrl pair.
+						 */
+						if (space < 2) {
+							no_space_left = 1;
+							break;
+						}
+						tty_put_char(tty, '^');
+						tty_put_char(tty, op ^ 0100);
+						tty->column += 2;
+						space -= 2;
+					} else {
+						if (!space) {
+							no_space_left = 1;
+							break;
+						}
+						tty_put_char(tty, op);
+						space--;
+					}
+				}
+				/*
+				 * If above falls through, this was an
+				 * undefined op.
+				 */
+				cp += 2;
+				nr -= 2;
+			}
+
+			if (no_space_left)
+				break;
+		} else {
+			int retval;
+
+			if ((retval = do_output_char(c, tty, space)) < 0)
+				break;
+			space -= retval;
+			cp += 1;
+			nr -= 1;
+		}
+
+		/* When end of circular buffer reached, wrap around */
+		if (cp >= buf_end)
+			cp -= N_TTY_BUF_SIZE;
+	}
+
+	if (nr == 0) {
+		tty->echo_pos = 0;
+		tty->echo_cnt = 0;
+		tty->echo_overrun = 0;
+	} else {
+		int num_processed = tty->echo_cnt - nr;
+		tty->echo_pos += num_processed;
+		tty->echo_pos &= N_TTY_BUF_SIZE - 1;
+		tty->echo_cnt = nr;
+		if (num_processed > 0)
+			tty->echo_overrun = 0;
+	}
+
+	mutex_unlock(&tty->echo_lock);
+	mutex_unlock(&tty->output_lock);
+
+	if (tty->ops->flush_chars)
+		tty->ops->flush_chars(tty);
+}
 
 /**
- *	echo_char	-	echo characters
+ *	add_echo_byte	-	add a byte to the echo buffer
+ *	@c: unicode byte to echo
+ *	@tty: terminal device
+ *
+ *	Add a character or operation byte to the echo buffer.
+ *
+ *	Should be called under the echo lock to protect the echo buffer.
+ */
+
+static void add_echo_byte(unsigned char c, struct tty_struct *tty)
+{
+	int	new_byte_pos;
+
+	if (tty->echo_cnt == N_TTY_BUF_SIZE) {
+		/* Circular buffer is already at capacity */
+		new_byte_pos = tty->echo_pos;
+
+		/*
+		 * Since the buffer start position needs to be advanced,
+		 * be sure to step by a whole operation byte group.
+		 */
+		if (tty->echo_buf[tty->echo_pos] == ECHO_OP_START)
+		{
+			if (tty->echo_buf[(tty->echo_pos + 1) &
+					  (N_TTY_BUF_SIZE - 1)] ==
+						ECHO_OP_ERASE_TAB) {
+				tty->echo_pos += 3;
+				tty->echo_cnt -= 2;
+			} else {
+				tty->echo_pos += 2;
+				tty->echo_cnt -= 1;
+			}
+		} else {
+			tty->echo_pos++;
+		}
+		tty->echo_pos &= N_TTY_BUF_SIZE - 1;
+
+		tty->echo_overrun = 1;
+	} else {
+		new_byte_pos = tty->echo_pos + tty->echo_cnt;
+		new_byte_pos &= N_TTY_BUF_SIZE - 1;
+		tty->echo_cnt++;
+	}
+
+	tty->echo_buf[new_byte_pos] = c;
+}
+
+/**
+ *	echo_move_back_col	-	add operation to move back a column
+ *	@tty: terminal device
+ *
+ *	Add an operation to the echo buffer to move back one column.
+ *
+ *	Locking: echo_lock to protect the echo buffer
+ */
+
+static void echo_move_back_col(struct tty_struct *tty)
+{
+	mutex_lock(&tty->echo_lock);
+
+	add_echo_byte(ECHO_OP_START, tty);
+	add_echo_byte(ECHO_OP_MOVE_BACK_COL, tty);
+
+	mutex_unlock(&tty->echo_lock);
+}
+
+/**
+ *	echo_set_canon_col	-	add operation to set the canon column
+ *	@tty: terminal device
+ *
+ *	Add an operation to the echo buffer to set the canon column
+ *	to the current column.
+ *
+ *	Locking: echo_lock to protect the echo buffer
+ */
+
+static void echo_set_canon_col(struct tty_struct *tty)
+{
+	mutex_lock(&tty->echo_lock);
+
+	add_echo_byte(ECHO_OP_START, tty);
+	add_echo_byte(ECHO_OP_SET_CANON_COL, tty);
+
+	mutex_unlock(&tty->echo_lock);
+}
+
+/**
+ *	echo_erase_tab	-	add operation to erase a tab
+ *	@num_chars: number of character columns already used
+ *	@after_tab: true if num_chars starts after a previous tab
+ *	@tty: terminal device
+ *
+ *	Add an operation to the echo buffer to erase a tab.
+ *
+ *	Called by the eraser function, which knows how many character
+ *	columns have been used since either a previous tab or the start
+ *	of input.  This information will be used later, along with
+ *	canon column (if applicable), to go back the correct number
+ *	of columns.
+ *
+ *	Locking: echo_lock to protect the echo buffer
+ */
+
+static void echo_erase_tab(unsigned int num_chars, int after_tab,
+			   struct tty_struct *tty)
+{
+	mutex_lock(&tty->echo_lock);
+
+	add_echo_byte(ECHO_OP_START, tty);
+	add_echo_byte(ECHO_OP_ERASE_TAB, tty);
+
+	/* We only need to know this modulo 8 (tab spacing) */
+	num_chars &= 7;
+
+	/* Set the high bit as a flag if num_chars is after a previous tab */
+	if (after_tab)
+		num_chars |= 0x80;
+	
+	add_echo_byte(num_chars, tty);
+
+	mutex_unlock(&tty->echo_lock);
+}
+
+/**
+ *	echo_char_raw	-	echo a character raw
  *	@c: unicode byte to echo
  *	@tty: terminal device
  *
  *	Echo user input back onto the screen. This must be called only when
  *	L_ECHO(tty) is true. Called from the driver receive_buf path.
  *
- *	Relies on BKL for tty column locking
+ *	This variant does not treat control characters specially.
+ *
+ *	Locking: echo_lock to protect the echo buffer
+ */
+
+static void echo_char_raw(unsigned char c, struct tty_struct *tty)
+{
+	mutex_lock(&tty->echo_lock);
+
+	if (c == ECHO_OP_START) {
+		add_echo_byte(ECHO_OP_START, tty);
+		add_echo_byte(ECHO_OP_START, tty);
+	} else {
+		add_echo_byte(c, tty);
+	}
+
+	mutex_unlock(&tty->echo_lock);
+}
+
+/**
+ *	echo_char	-	echo a character
+ *	@c: unicode byte to echo
+ *	@tty: terminal device
+ *
+ *	Echo user input back onto the screen. This must be called only when
+ *	L_ECHO(tty) is true. Called from the driver receive_buf path.
+ *
+ *	This variant tags control characters to be possibly echoed as
+ *	as "^X" (where X is the letter representing the control char).
+ *
+ *	Locking: echo_lock to protect the echo buffer
  */
 
 static void echo_char(unsigned char c, struct tty_struct *tty)
 {
-	if (L_ECHOCTL(tty) && iscntrl(c) && c != '\t') {
-		tty_put_char(tty, '^');
-		tty_put_char(tty, c ^ 0100);
-		tty->column += 2;
-	} else
-		opost(c, tty);
+	mutex_lock(&tty->echo_lock);
+
+	if (c == ECHO_OP_START) {
+		add_echo_byte(ECHO_OP_START, tty);
+		add_echo_byte(ECHO_OP_START, tty);
+	} else {
+		if (iscntrl(c) && c != '\t')
+			add_echo_byte(ECHO_OP_START, tty);
+		add_echo_byte(c, tty);
+	}
+
+	mutex_unlock(&tty->echo_lock);
 }
 
 /**
- *	finsh_erasing		-	complete erase
+ *	finish_erasing		-	complete erase
  *	@tty: tty doing the erase
- *
- *	Relies on BKL for tty column locking
  */
+
 static inline void finish_erasing(struct tty_struct *tty)
 {
 	if (tty->erasing) {
-		tty_put_char(tty, '/');
-		tty->column++;
+		echo_char_raw('/', tty);
 		tty->erasing = 0;
 	}
 }
@@ -460,7 +856,7 @@
  *	present in the stream from the driver layer. Handles the complexities
  *	of UTF-8 multibyte symbols.
  *
- *	Locking: read_lock for tty buffers, BKL for column/erasing state
+ *	Locking: read_lock for tty buffers
  */
 
 static void eraser(unsigned char c, struct tty_struct *tty)
@@ -471,7 +867,7 @@
 
 	/* FIXME: locking needed ? */
 	if (tty->read_head == tty->canon_head) {
-		/* opost('\a', tty); */		/* what do you think? */
+		/* echo_char_raw('\a', tty); */ /* what do you think? */
 		return;
 	}
 	if (c == ERASE_CHAR(tty))
@@ -497,7 +893,7 @@
 			echo_char(KILL_CHAR(tty), tty);
 			/* Add a newline if ECHOK is on and ECHOKE is off. */
 			if (L_ECHOK(tty))
-				opost('\n', tty);
+				echo_char_raw('\n', tty);
 			return;
 		}
 		kill_type = KILL;
@@ -533,67 +929,62 @@
 		if (L_ECHO(tty)) {
 			if (L_ECHOPRT(tty)) {
 				if (!tty->erasing) {
-					tty_put_char(tty, '\\');
-					tty->column++;
+					echo_char_raw('\\', tty);
 					tty->erasing = 1;
 				}
 				/* if cnt > 1, output a multi-byte character */
 				echo_char(c, tty);
 				while (--cnt > 0) {
 					head = (head+1) & (N_TTY_BUF_SIZE-1);
-					tty_put_char(tty, tty->read_buf[head]);
+					echo_char_raw(tty->read_buf[head], tty);
+					echo_move_back_col(tty);
 				}
 			} else if (kill_type == ERASE && !L_ECHOE(tty)) {
 				echo_char(ERASE_CHAR(tty), tty);
 			} else if (c == '\t') {
-				unsigned int col = tty->canon_column;
-				unsigned long tail = tty->canon_head;
+				unsigned int num_chars = 0;
+				int after_tab = 0;
+				unsigned long tail = tty->read_head;
 
-				/* Find the column of the last char. */
-				while (tail != tty->read_head) {
+				/*
+				 * Count the columns used for characters
+				 * since the start of input or after a
+				 * previous tab.
+				 * This info is used to go back the correct
+				 * number of columns.
+				 */
+				while (tail != tty->canon_head) {
+					tail = (tail-1) & (N_TTY_BUF_SIZE-1);
 					c = tty->read_buf[tail];
-					if (c == '\t')
-						col = (col | 7) + 1;
+					if (c == '\t') {
+						after_tab = 1;
+						break;
+					}
 					else if (iscntrl(c)) {
 						if (L_ECHOCTL(tty))
-							col += 2;
-					} else if (!is_continuation(c, tty))
-						col++;
-					tail = (tail+1) & (N_TTY_BUF_SIZE-1);
+							num_chars += 2;
+					} else if (!is_continuation(c, tty)) {
+						num_chars++;
+					}
 				}
-
-				/* should never happen */
-				if (tty->column > 0x80000000)
-					tty->column = 0;
-
-				/* Now backup to that column. */
-				while (tty->column > col) {
-					/* Can't use opost here. */
-					tty_put_char(tty, '\b');
-					if (tty->column > 0)
-						tty->column--;
-				}
+				echo_erase_tab(num_chars, after_tab, tty);
 			} else {
 				if (iscntrl(c) && L_ECHOCTL(tty)) {
-					tty_put_char(tty, '\b');
-					tty_put_char(tty, ' ');
-					tty_put_char(tty, '\b');
-					if (tty->column > 0)
-						tty->column--;
+					echo_char_raw('\b', tty);
+					echo_char_raw(' ', tty);
+					echo_char_raw('\b', tty);
 				}
 				if (!iscntrl(c) || L_ECHOCTL(tty)) {
-					tty_put_char(tty, '\b');
-					tty_put_char(tty, ' ');
-					tty_put_char(tty, '\b');
-					if (tty->column > 0)
-						tty->column--;
+					echo_char_raw('\b', tty);
+					echo_char_raw(' ', tty);
+					echo_char_raw('\b', tty);
 				}
 			}
 		}
 		if (kill_type == ERASE)
 			break;
 	}
-	if (tty->read_head == tty->canon_head)
+	if (tty->read_head == tty->canon_head && L_ECHO(tty))
 		finish_erasing(tty);
 }
 
@@ -724,14 +1115,18 @@
 		c=tolower(c);
 
 	if (tty->stopped && !tty->flow_stopped && I_IXON(tty) &&
-	    ((I_IXANY(tty) && c != START_CHAR(tty) && c != STOP_CHAR(tty)) ||
-	     c == INTR_CHAR(tty) || c == QUIT_CHAR(tty) || c == SUSP_CHAR(tty)))
+	    I_IXANY(tty) && c != START_CHAR(tty) && c != STOP_CHAR(tty) &&
+	    c != INTR_CHAR(tty) && c != QUIT_CHAR(tty) && c != SUSP_CHAR(tty)) {
 		start_tty(tty);
+		process_echoes(tty);
+	}
 
 	if (tty->closing) {
 		if (I_IXON(tty)) {
-			if (c == START_CHAR(tty))
+			if (c == START_CHAR(tty)) {
 				start_tty(tty);
+				process_echoes(tty);
+			}
 			else if (c == STOP_CHAR(tty))
 				stop_tty(tty);
 		}
@@ -745,17 +1140,20 @@
 	 * up.
 	 */
 	if (!test_bit(c, tty->process_char_map) || tty->lnext) {
-		finish_erasing(tty);
 		tty->lnext = 0;
 		if (L_ECHO(tty)) {
+			finish_erasing(tty);
 			if (tty->read_cnt >= N_TTY_BUF_SIZE-1) {
-				tty_put_char(tty, '\a'); /* beep if no space */
+				/* beep if no space */
+				echo_char_raw('\a', tty);
+				process_echoes(tty);
 				return;
 			}
 			/* Record the column of first canon char. */
 			if (tty->canon_head == tty->read_head)
-				tty->canon_column = tty->column;
+				echo_set_canon_col(tty);
 			echo_char(c, tty);
+			process_echoes(tty);
 		}
 		if (I_PARMRK(tty) && c == (unsigned char) '\377')
 			put_tty_queue(c, tty);
@@ -766,6 +1164,7 @@
 	if (I_IXON(tty)) {
 		if (c == START_CHAR(tty)) {
 			start_tty(tty);
+			process_echoes(tty);
 			return;
 		}
 		if (c == STOP_CHAR(tty)) {
@@ -786,7 +1185,6 @@
 		if (c == SUSP_CHAR(tty)) {
 send_signal:
 			/*
-			 * Echo character, and then send the signal.
 			 * Note that we do not use isig() here because we want
 			 * the order to be:
 			 * 1) flush, 2) echo, 3) signal
@@ -795,8 +1193,12 @@
 				n_tty_flush_buffer(tty);
 				tty_driver_flush_buffer(tty);
 			}
-			if (L_ECHO(tty))
+			if (I_IXON(tty))
+				start_tty(tty);
+			if (L_ECHO(tty)) {
 				echo_char(c, tty);
+				process_echoes(tty);
+			}
 			if (tty->pgrp)
 				kill_pgrp(tty->pgrp, signal, 1);
 			return;
@@ -815,6 +1217,7 @@
 		if (c == ERASE_CHAR(tty) || c == KILL_CHAR(tty) ||
 		    (c == WERASE_CHAR(tty) && L_IEXTEN(tty))) {
 			eraser(c, tty);
+			process_echoes(tty);
 			return;
 		}
 		if (c == LNEXT_CHAR(tty) && L_IEXTEN(tty)) {
@@ -822,8 +1225,9 @@
 			if (L_ECHO(tty)) {
 				finish_erasing(tty);
 				if (L_ECHOCTL(tty)) {
-					tty_put_char(tty, '^');
-					tty_put_char(tty, '\b');
+					echo_char_raw('^', tty);
+					echo_char_raw('\b', tty);
+					process_echoes(tty);
 				}
 			}
 			return;
@@ -834,18 +1238,20 @@
 
 			finish_erasing(tty);
 			echo_char(c, tty);
-			opost('\n', tty);
+			echo_char_raw('\n', tty);
 			while (tail != tty->read_head) {
 				echo_char(tty->read_buf[tail], tty);
 				tail = (tail+1) & (N_TTY_BUF_SIZE-1);
 			}
+			process_echoes(tty);
 			return;
 		}
 		if (c == '\n') {
 			if (L_ECHO(tty) || L_ECHONL(tty)) {
 				if (tty->read_cnt >= N_TTY_BUF_SIZE-1)
-					tty_put_char(tty, '\a');
-				opost('\n', tty);
+					echo_char_raw('\a', tty);
+				echo_char_raw('\n', tty);
+				process_echoes(tty);
 			}
 			goto handle_newline;
 		}
@@ -862,11 +1268,12 @@
 			 */
 			if (L_ECHO(tty)) {
 				if (tty->read_cnt >= N_TTY_BUF_SIZE-1)
-					tty_put_char(tty, '\a');
+					echo_char_raw('\a', tty);
 				/* Record the column of first canon char. */
 				if (tty->canon_head == tty->read_head)
-					tty->canon_column = tty->column;
+					echo_set_canon_col(tty);
 				echo_char(c, tty);
+				process_echoes(tty);
 			}
 			/*
 			 * XXX does PARMRK doubling happen for
@@ -889,20 +1296,23 @@
 		}
 	}
 
-	finish_erasing(tty);
 	if (L_ECHO(tty)) {
+		finish_erasing(tty);
 		if (tty->read_cnt >= N_TTY_BUF_SIZE-1) {
-			tty_put_char(tty, '\a'); /* beep if no space */
+			/* beep if no space */
+			echo_char_raw('\a', tty);
+			process_echoes(tty);
 			return;
 		}
 		if (c == '\n')
-			opost('\n', tty);
+			echo_char_raw('\n', tty);
 		else {
 			/* Record the column of first canon char. */
 			if (tty->canon_head == tty->read_head)
-				tty->canon_column = tty->column;
+				echo_set_canon_col(tty);
 			echo_char(c, tty);
 		}
+		process_echoes(tty);
 	}
 
 	if (I_PARMRK(tty) && c == (unsigned char) '\377')
@@ -923,6 +1333,9 @@
 
 static void n_tty_write_wakeup(struct tty_struct *tty)
 {
+	/* Write out any echoed characters that are still pending */
+	process_echoes(tty);
+	
 	if (tty->fasync) {
 		set_bit(TTY_DO_WRITE_WAKEUP, &tty->flags);
 		kill_fasync(&tty->fasync, SIGIO, POLL_OUT);
@@ -1134,6 +1547,10 @@
 		free_buf(tty->read_buf);
 		tty->read_buf = NULL;
 	}
+	if (tty->echo_buf) {
+		free_buf(tty->echo_buf);
+		tty->echo_buf = NULL;
+	}
 }
 
 /**
@@ -1151,13 +1568,19 @@
 	if (!tty)
 		return -EINVAL;
 
-	/* This one is ugly. Currently a malloc failure here can panic */
+	/* These are ugly. Currently a malloc failure here can panic */
 	if (!tty->read_buf) {
 		tty->read_buf = alloc_buf();
 		if (!tty->read_buf)
 			return -ENOMEM;
 	}
+	if (!tty->echo_buf) {
+		tty->echo_buf = alloc_buf();
+		if (!tty->echo_buf)
+			return -ENOMEM;
+	}
 	memset(tty->read_buf, 0, N_TTY_BUF_SIZE);
+	memset(tty->echo_buf, 0, N_TTY_BUF_SIZE);
 	reset_buffer_flags(tty);
 	tty->column = 0;
 	n_tty_set_termios(tty, NULL);
@@ -1487,16 +1910,23 @@
  *	@buf: userspace buffer pointer
  *	@nr: size of I/O
  *
- *	Write function of the terminal device. This is serialized with
+ *	Write function of the terminal device.  This is serialized with
  *	respect to other write callers but not to termios changes, reads
- *	and other such events. We must be careful with N_TTY as the receive
- *	code will echo characters, thus calling driver write methods.
+ *	and other such events.  Since the receive code will echo characters,
+ *	thus calling driver write methods, the output_lock is used in
+ *	the output processing functions called here as well as in the
+ *	echo processing function to protect the column state and space
+ *	left in the buffer.
  *
  *	This code must be sure never to sleep through a hangup.
+ *
+ *	Locking: output_lock to protect column state and space left
+ *		 (note that the process_output*() functions take this
+ *		  lock themselves)
  */
 
 static ssize_t n_tty_write(struct tty_struct *tty, struct file *file,
-			  const unsigned char *buf, size_t nr)
+			   const unsigned char *buf, size_t nr)
 {
 	const unsigned char *b = buf;
 	DECLARE_WAITQUEUE(wait, current);
@@ -1510,6 +1940,9 @@
 			return retval;
 	}
 
+	/* Write out any echoed characters that are still pending */
+	process_echoes(tty);
+	
 	add_wait_queue(&tty->write_wait, &wait);
 	while (1) {
 		set_current_state(TASK_INTERRUPTIBLE);
@@ -1523,7 +1956,7 @@
 		}
 		if (O_OPOST(tty) && !(test_bit(TTY_HW_COOK_OUT, &tty->flags))) {
 			while (nr > 0) {
-				ssize_t num = opost_block(tty, b, nr);
+				ssize_t num = process_output_block(tty, b, nr);
 				if (num < 0) {
 					if (num == -EAGAIN)
 						break;
@@ -1535,7 +1968,7 @@
 				if (nr == 0)
 					break;
 				c = *b;
-				if (opost(c, tty) < 0)
+				if (process_output(c, tty) < 0)
 					break;
 				b++; nr--;
 			}
@@ -1663,4 +2096,3 @@
 	.receive_buf     = n_tty_receive_buf,
 	.write_wakeup    = n_tty_write_wakeup
 };
-