V4L/DVB (10854): cx18: Correct comments about vertical and horizontal blanking timings

This change only affects comments.
Signed-off-by: Andy Walls <awalls@radix.net>
Signed-off-by: Mauro Carvalho Chehab <mchehab@redhat.com>

drivers/media/video/cx18/cx18-av-core.c

@@ -292,23 +292,29 @@ void cx18_av_std_setup(struct cx18 *cx)
 		 *
 		 * vsync:     always 6 half-lines of vsync pulses
 		 * vactive:   half lines of active video
-		 * vblank656: half lines, after line 3, of blanked video
-		 * vblank:    half lines, after line 9, of blanked video
+		 * vblank656: half lines, after line 3/mid-266, of blanked video
+		 * vblank:    half lines, after line 9/272, of blanked video
 		 *
+		 * As far as I can tell:
 		 * vblank656 starts counting from the falling edge of the first
-		 * 	vsync pulse (start of line 4)
+		 * 	vsync pulse (start of line 4 or mid-266)
 		 * vblank starts counting from the after the 6 vsync pulses and
-		 * 	6 equalization pulses (start of line 10)
+		 * 	6 or 5 equalization pulses (start of line 10 or 272)
 		 *
 		 * For 525 line systems the driver will extract VBI information
-		 * from lines 10 through 21.  To avoid the EAV RP code from
-		 * toggling at the start of hblank at line 22, where sliced VBI
-		 * data from line 21 is stuffed, also treat line 22 as blanked.
+		 * from lines 10-21 and lines 273-284.
 		 */
-		vblank656 = 38; /* lines  4 through  22 */
-		vblank = 26;	/* lines 10 through  22 */
-		vactive = 481;  /* lines 23 through 262.5 */
+		vblank656 = 38; /* lines  4 -  22  &  266 - 284 */
+		vblank = 26;	/* lines 10 -  22  &  272 - 284 */
+		vactive = 481;  /* lines 23 - 263  &  285 - 525 */
 
+		/*
+		 * For a 13.5 Mpps clock and 15,734.26 Hz line rate, a line is
+		 * is 858 pixels = 720 active + 138 blanking.  The Hsync leading
+		 * edge should happen 1.2 us * 13.5 Mpps ~= 16 pixels after the
+		 * end of active video, leaving 122 pixels of hblank to ignore
+		 * before active video starts.
+		 */
 		hactive = 720;
 		hblank = 122;
 		luma_lpf = 1;

drivers/media/video/cx18/cx18-av-core.h

@@ -89,16 +89,21 @@ struct cx18_av_state {
 
 	/*
 	 * The VBI slicer starts operating and counting lines, begining at
-	 * slicer line count of 1, at D lines after the deassertion of VRESET
-	 * This staring field line, S, is 6 or 10 for 625 or 525 line systems.
-	 * Sliced ancillary data captured on VBI slicer line M is sent at the
-	 * beginning of the next VBI slicer line, VBI slicer line count N = M+1.
-	 * Thus when the VBI slicer reports a VBI slicer line number with
-	 * ancillary data, the IDID0 byte indicates VBI slicer line N.
-	 * The actual field line that the captured data comes from is
+	 * slicer line count of 1, at D lines after the deassertion of VRESET.
+	 * This staring field line, S, is 6 (& 319) or 10 (& 273) for 625 or 525
+	 * line systems respectively.  Sliced ancillary data captured on VBI
+	 * slicer line M is inserted after the VBI slicer is done with line M,
+	 * when VBI slicer line count is N = M+1.  Thus when the VBI slicer
+	 * reports a VBI slicer line number with ancillary data, the IDID0 byte
+	 * indicates VBI slicer line N.  The actual field line that the captured
+	 * data comes from is
+	 *
 	 * L = M+(S+D-1) = N-1+(S+D-1) = N + (S+D-2).
 	 *
+	 * L is the line in the field, not frame, from which the VBI data came.
+	 * N is the line reported by the slicer in the ancillary data.
 	 * D is the slicer_line_delay value programmed into register 0x47f.
+	 * S is 6 for 625 line systems or 10 for 525 line systems
 	 * (S+D-2) is the slicer_line_offset used to convert slicer reported
 	 * line counts to actual field lines.
 	 */

drivers/media/video/cx18/cx18-streams.c

@@ -413,9 +413,8 @@ static void cx18_vbi_setup(struct cx18_stream *s)
 		 * 0x90 (Task                         HorizontalBlank)
 		 * 0xd0 (Task EvenField               HorizontalBlank)
 		 *
-		 * We have set the digitzer to consider the first active line
-		 * as part of VerticalBlank as well so we don't have to look for
-		 * these problem codes nor lose the last line of sliced data.
+		 * We have set the digitzer such that we don't have to worry
+		 * about these problem codes.
 		 */
 		data[4] = 0xB0F0B0F0;
 		/*