2003-03-26 10:42:32

by Petr Vandrovec

[permalink] [raw]
Subject: Re: [Linux-fbdev-devel] [BK FBDEV] A few more updates.

On 26 Mar 03 at 11:42, Petr Vandrovec wrote:
>
> accel_cursor unconditionally sets FB_CUR_SETPOS. Can you write it
> down to the TODO list to eliminate this? Cursor position lives
> in different registers than cursor enable/disable on my hardware...

> And if we could rename FB_CUR_SETCUR to FB_CUR_SETVISIBILITY and
> leave cursor->enable setting on accel_cursor's caller, it would
> be even better.

I just noticed that softcursor.c contains copy of FB_CUR_SETCUR ->
cursor->enable code from accel_cursor(). Either 'enable' field has
no bussiness in the fb_cursor structure (as it can be always infered from
set flags, and if softcursor needs some internal bookkeeping, it should
use some other variable and not fb_cursor's field), or this conversion
should disappear from softcursor.c, and FB_CUR_SETCUR semantic should
change.
Petr