Remove code for skipping duplicate port 80 writes

The kernel no longer uses port 80 as a delay mechanism, so we don't
need to detect the no-longer-present spammy writes.

Signed-off-by: Randall Spangler <rspangler@chromium.org>

BUG=chrome-os-partner:7972
TEST=port80 scroll, then boot the system.  see a few repeated bytes,
but not piles of 00 and ff's.

Change-Id: Id14dc43ab4e1b15c6bab99a17c062f295a59e7e6

board/link/board.h

@@ -55,11 +55,6 @@
 /* LPC COMx I/O address (in x86 I/O address space) */
 #define LPC_COMX_ADDR 0x3f8  /* COM1 */
 
-/* Define the following to print repeated duplicate port 80 writes.  Normally
- * we only print the port 80 value when it changes, because the kernel spams
- * port 80 with repeated writes as a delay mechanism. */
-#define CONFIG_PORT80_PRINT_DUPLICATES
-
 /* ADC inputs */
 /* TODO: assign real ADC inputs */
 #define ADC_IN0 11  /* Charger current */

common/port80.c

@@ -20,19 +20,10 @@ static int scroll = 0;
 
 void port_80_write(int data)
 {
-#ifndef CONFIG_PORT80_PRINT_DUPLICATES
-	static int last_data = -1;  /* Last data written to port 80 */
-
-	/* Ignore duplicate writes, since the linux kernel writes to port 80
-	 * as a delay mechanism during boot. */
-	if (data == last_data)
-		return;
-
-	last_data = data;
-#endif
-
-	/* TODO: post to SWI and print from there?  This currently
-	 * prints from inside the LPC interrupt itself. */
+	/* Note that this currently prints from inside the LPC interrupt
+	 * itself.  Probably not worth the system overhead to buffer the data
+	 * and print it from a task, because we're printing a small amount of
+	 * data and uart_printf() doesn't block. */
 	uart_printf("%c[Port 80: 0x%02x]", scroll ? '\n' : '\r', data);
 
 	history[head] = data;