endian problem ?

endian problem ?

[Benjamin Herrenschmidt]

Ok, I've ported radeonfb, controlfb and platinumfb to the
new API (which is quite nice since allowed me to remove
a whole bunch of cruft from these drivers, good job !)

However, I'm having a problem: With controlfb and
radeonfb (couldn't test platinum yet but it's very similar
to controlfb), the console is fine in 8 bpp, but text is
plain blue in any higher bit depth (15, 16 or 32).

X with "fbdev" driver works fine, so I think setcolreg
works properly. All those drivers are set to DIRECTCOLOR
for these bit depth. Maybe I've missed something about
the pseudo_palette thing (is it documented somewhere ?)
or are there known endian bugs in the cfb_* routines ?

Ben.




-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf

Re: endian problem ?

[James Simmons]

On Sun, 13 Oct 2002, Benjamin Herrenschmidt wrote:

> Ok, I've ported radeonfb, controlfb and platinumfb to the
> new API (which is quite nice since allowed me to remove
> a whole bunch of cruft from these drivers, good job !)

Great!!!! Just let me know when you want them placed into fbdev BK.

> However, I'm having a problem: With controlfb and
> radeonfb (couldn't test platinum yet but it's very similar
> to controlfb), the console is fine in 8 bpp, but text is
> plain blue in any higher bit depth (15, 16 or 32).

Okay I have seen this problem on a few drivers. It is either a bug in
setcolreg which is often copied from one place to another or a bug in the
upper levels. I seen this on the atyfb driver happen.

> X with "fbdev" driver works fine, so I think setcolreg
> works properly. All those drivers are set to DIRECTCOLOR
> for these bit depth. Maybe I've missed something about
> the pseudo_palette thing (is it documented somewhere ?)
> or are there known endian bugs in the cfb_* routines ?

I now know it isn't a endian issues since on my little endian machine I
have the same problem. I don't know if it is a cfb_* bug. I will try out
my tdfx card to see. It has accelerated functions so the error woudln't be
there.

MS: (n) 1. A debilitating and surprisingly widespread affliction that
renders the sufferer barely able to perform the simplest task. 2. A disease.

James Simmons  [jsimmons@users.sf.net] 	                ____/|
fbdev/console/gfx developer                             \ o.O|
http://www.linux-fbdev.org                               =(_)=
http://linuxgfx.sourceforge.net                            U
http://linuxconsole.sourceforge.net



-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf