[PATCH] inivisible cursor on vesafb

Hi Linus,

Since vesafb was converted to the (strange) cfb* interface, it uses random
colors for the cursor, making cursor invisible when console uses black foreground
(f.e. Midnight Commander does that). Problem is common to all fbdevs which
use fbcon-accel interface.

This change tries to restore old behavior which existed before cfb* came in.
Unfortunately due to region.color semantic it is not possible to completely
restore previous behavior - 
+ for truecolor cfbfillrect always consults current palette, so we do not 
  have "always invert" constant available, and so we do inversion with palette
  color 15 (it is light white, ~0, by default)
+ for directcolor it assumes that display has at least 4 bits per color.
+ for pseudocolor it works as old code worked.

						Thanks,
							Petr Vandrovec
							vandrove@vc.cvut.cz


diff -urdN linux/drivers/video/fbcon-accel.c linux/drivers/video/fbcon-accel.c

--- linux/drivers/video/fbcon-accel.c	2002-08-18 01:35:06.000000000 +0200
+++ linux/drivers/video/fbcon-accel.c	2002-08-18 15:07:42.000000000 +0200

@@ -104,7 +104,10 @@
 	struct fb_info *info = p->fb_info;
 	struct fb_fillrect region;
 
-	region.color = attr_fgcol_ec(p, p->conp);
+	if (info->fix.visual == FB_VISUAL_PSEUDOCOLOR || info->fix.visual == FB_VISUAL_TRUECOLOR)
+		region.color = 0x0F;
+	else
+		region.color = (0x0F << info->var.red.offset) | (0x0F << info->var.green.offset) | (0x0F << info->var.blue.offset);
 	region.dx = xx * fontwidth(p);
 	region.dy = yy * fontheight(p);
 	region.width  = fontwidth(p);


-------------------------------------------------------

This sf.net email is sponsored by: OSDN - Tired of that same old
cell phone?  Get a new here for FREE!
https://www.inphonic.com/r.asp?r=sourceforge1&refcode1=vs3390

On Mon, 2002-08-19 at 07:05, Petr Vandrovec wrote:

Hi Linus,

Since vesafb was converted to the (strange) cfb* interface, it uses random
colors for the cursor, making cursor invisible when console uses black foreground
(f.e. Midnight Commander does that). Problem is common to all fbdevs which
use fbcon-accel interface.

This change tries to restore old behavior which existed before cfb* came in.
Unfortunately due to region.color semantic it is not possible to completely
restore previous behavior - 
+ for truecolor cfbfillrect always consults current palette, so we do not 
  have "always invert" constant available, and so we do inversion with palette
  color 15 (it is light white, ~0, by default)
+ for directcolor it assumes that display has at least 4 bits per color.
+ for pseudocolor it works as old code worked.

I think it has been agreed that all color entries in fbcon-accel should
always be an index to the pseudopalette regardless of the visuals.  So a
more consistent change is to have cfbfillrect look up the color from the
pseudopalette assuming that the driver correctly loaded the value to the
pseudopalette.