Re: [PATCH 1/1 2.6.15-rc4-git1] Fix switching to KD_TEXT

["Antonino A. Daplas" <adaplas@gmail.com>]

Knut Petersen wrote:
>  Every framebuffer driver relies on the assumption that the set_par()
> function
> of the driver is called before drawing functions and other functions
> dependent
> on the hardware state are executed.
> 
> This assumption is false in two cases, and one is a genuine linux
> bug:
> 
>    1: Whenever you switch from X to a framebuffer console for the very
>        first time, there is a chance that a broken X system has _not_ set
>        the mode to KD_GRAPHICS, thus the vt and framebuffer code
>        executes a screen redraw and several other functions before a
>        set_par() is executed. This is believed to be not a bug of linux
>        but a bug of X/xdm.
> 
>    2: Whenever a switch from X to a framebuffer console occures,
>         the pan_display() function of the driver is called once before
>         the set_par() function of the driver is called. Code path:
>         complete_change_console -> redraw_screen -> fbcon_switch ->
>         bit_update_start-> fb_pan_display -> xyz_pan_display.
>         This is clearly a bug of linux.

This part, #2, can be easily fixed.

> 
> Although our primary goal must be to fix linux bugs and not to work
> around bugs of X, the patch fixes both of the cases.
> 
> The advantage and correctness of this patch should be obvious. Yes, it
> does add a possibly slow call to the fb_set_par() function, but at this
> point it is necessary.
> 
> Signed-off-by: Knut Petersen <Knut_Petersen@t-online.de>

Sorry, NAK for now.  Unless other people agree that it is okay for them
to have an unconditional call to set_par() for every console switch. Note
that the set_par() in some drivers is terribly slow (several seconds at least).

Let's wait a few days, if nobody disagrees with you, so be it.

Tony