[PATCH 2.6][RESEND] fbcon margins colour fix

[PATCH 2.6][RESEND] fbcon margins colour fix

[Jakub Bogusz]

[-- Attachment #1: Type: text/plain, Size: 260 bytes --]

I sent it a few times to linux-kernel and at least one to
linux-fbdev-devel, but haven't seen any comments - and this annoying
changing margins colour seems to be still there in 2.6.4 (at least on
tdfxfb).


-- 
Jakub Bogusz    http://cyber.cs.net.pl/~qboosh/

[-- Attachment #2: linux-fbcon-margins.patch --]
[-- Type: text/plain, Size: 1373 bytes --]

This fixes "margin colour" (colour used to clear top and bottom margins -
e.g. remaining half of line at the bottom of 100x37 console on 800x600
framebuffer).

I don't know what was the intention behind using attr_bgcol_ec() here,
but it caused using of background colour of last erase character to clear
margins (thus it was changing from time to time when switching consoles
or so).
I think that margin colour should be equal overscan colour, which seems
to be always black.
This patch changes margin colour to black (or colour 0 in palette modes)
in accel_clear_margins() and removes unused then bgshift variable.

	-- Jakub Bogusz <qboosh@pld-linux.org>

--- linux-2.6.4/drivers/video/console/fbcon.c.orig	2004-03-11 03:55:36.000000000 +0100
+++ linux-2.6.4/drivers/video/console/fbcon.c	2004-03-17 20:47:38.749388312 +0100
@@ -489,7 +489,6 @@
 void accel_clear_margins(struct vc_data *vc, struct fb_info *info,
 				int bottom_only)
 {
-	int bgshift = (vc->vc_hi_font_mask) ? 13 : 12;
 	unsigned int cw = vc->vc_font.width;
 	unsigned int ch = vc->vc_font.height;
 	unsigned int rw = info->var.xres - (vc->vc_cols*cw);
@@ -498,7 +497,7 @@
 	unsigned int bs = info->var.yres - bh;
 	struct fb_fillrect region;
 
-	region.color = attr_bgcol_ec(bgshift, vc);
+	region.color = 0; /* margins color = overscan color */
 	region.rop = ROP_COPY;
 
 	if (rw && !bottom_only) {

Re: [Linux-fbdev-devel] [PATCH 2.6][RESEND] fbcon margins colour fix

[Geert Uytterhoeven]

On Thu, 18 Mar 2004, Jakub Bogusz wrote:
> I sent it a few times to linux-kernel and at least one to
> linux-fbdev-devel, but haven't seen any comments - and this annoying
> changing margins colour seems to be still there in 2.6.4 (at least on
> tdfxfb).

What happens on `reverse video' (i.e. black on white, like Sun) graphics cards?
In that case the overscan color is white.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds

Re: [Linux-fbdev-devel] [PATCH 2.6][RESEND] fbcon margins colour fix

[Jakub Bogusz]

On Thu, Mar 18, 2004 at 10:21:31AM +0100, Geert Uytterhoeven wrote:
> On Thu, 18 Mar 2004, Jakub Bogusz wrote:
> > I sent it a few times to linux-kernel and at least one to
> > linux-fbdev-devel, but haven't seen any comments - and this annoying
> > changing margins colour seems to be still there in 2.6.4 (at least on
> > tdfxfb).
> 
> What happens on `reverse video' (i.e. black on white, like Sun) graphics cards?
> In that case the overscan color is white.

Uhm. What is palette entry for this white?
Or, more generally, how to find (palette colour number of) overscan
colour for current console?
Video erase character colour is not proper one as it may be different
even from background colour at the moment of vt switch.



-- 
Jakub Bogusz    http://cyber.cs.net.pl/~qboosh/

Re: [Linux-fbdev-devel] [PATCH 2.6][RESEND] fbcon margins colour fix

[Geert Uytterhoeven]

On Thu, 18 Mar 2004, Jakub Bogusz wrote:
> On Thu, Mar 18, 2004 at 10:21:31AM +0100, Geert Uytterhoeven wrote:
> > On Thu, 18 Mar 2004, Jakub Bogusz wrote:
> > > I sent it a few times to linux-kernel and at least one to
> > > linux-fbdev-devel, but haven't seen any comments - and this annoying
> > > changing margins colour seems to be still there in 2.6.4 (at least on
> > > tdfxfb).
> >
> > What happens on `reverse video' (i.e. black on white, like Sun) graphics cards?
> > In that case the overscan color is white.
>
> Uhm. What is palette entry for this white?

0, on pseudocolor displays. So probably it's OK.

> Or, more generally, how to find (palette colour number of) overscan
> colour for current console?
> Video erase character colour is not proper one as it may be different
> even from background colour at the moment of vt switch.

There's no standardized way for that. So it may depend on the driver. I know
ATI Mach64 has a special register for the overscan color, which is always set
to black, IIRC.

Gr{oetje,eeting}s,

						Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
							    -- Linus Torvalds