Subject: Re: VGA driver changes for machines without built-in font
To: Izumi Tsutsui <email@example.com>
From: Bang Jun-Young <firstname.lastname@example.org>
Date: 01/28/2003 15:23:14
On Sat, Jan 25, 2003 at 02:39:10AM +0900, Izumi Tsutsui wrote:
> Hmm. My intention of these changes is to use encoding
> of default wsfont in vga_no_builtinfont case.
> Without these changes, if the kernel uses wsfonts for default font
> other than WSDISPLAY_FONTENC_IBM, it shows some garbages.
> (for example, it fills screen with '@' (0x40) instead ' ' (0x20)
> on initialization with FONT_VT220ISO8x16, whose firstchar is 0x20.)
> I guess these implecit FONTENC_IBMs are used for the encoding of
> the builtin font (which is default), but is it wrong?
> In my patch, scr->encoding is set in vga_raster_setup_font().
Uhm, I wasn't right about that. Your code was right.
> > vga_reset() and its friends are not necessary for i386. How about making
> > it optionally compile via 'options VGA_RESET_ON_BOOT' (or something with
> > a better name)?
> Yeah, I also wonder if it should be optional or not..
> At the first time I added 'options VGA_NO_BUILTIN_FONT'
> to disable all related changes, but I removed it because
> it was also useful to use optional fonts even on i386.
> But actually vga_reset() is required by only some machines,
> so I'll add options VGA_RESET to enable it.
> (or needs a better name? :-)
Bang Jun-Young <email@example.com>