Tobias Hunger [EMAIL PROTECTED] writes:
Hello out there!
Guess what: My fbdevstill flickers, but I now think it is supposed to do so:-)
I had to switch monitors and got a better one then mine at home. So I
switched to fbdev (modprobe i2c-matroxfb), set a decent resolution (fbset
- - The flickering is independent of the refresh rate of the screen.
60Hz is as bad as 90Hz.
That pretty much rules out the "unsynchronized pageflipping" theory.
- - The flickering depends on the screen resolution! At 640x480 it is
hard to notice, at 1024x768 it awful.
This sounds like
On Fri, Jun 23, 2000 at 10:13:05AM +0200, Andreas Beck wrote:
On the pageflip demo:
Do you see both screens correctly,
or is one of them blank ?
Incidentally I discovered that on fbdev (kernel 2.2.11, matroxfb,
Millenium I), only the first page is visible, all other pages (e.g. F2
to F6 at
On Fri, Jun 23, 2000 at 03:26:52PM -0400, Steve Cheng wrote:
Incidentally I discovered that on fbdev (kernel 2.2.11, matroxfb,
Millenium I), only the first page is visible, all other pages (e.g. F2
to F6 at 640x480[GT_8BIT]) are blank when switched to. Set mode,
setwriteframe seems ok.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Wed, 21 Jun 2000, Marcus Sundberg wrote:
That's not double buffering, that's just, well, copying.
Doublebuffering is when you have one buffer being drawn to, and
another being displayed, and then swap between them when drawing is
finished.
Stefan Seefeld [EMAIL PROTECTED] writes:
Andrew Apted wrote:
If Berlin can do double buffering, then that might fix the problem.
that's what we do anyway (as Tobias states above). libart draws into a
memory visual which then gets copied with ggiCrossBlit.
That's not double buffering,
On Mon, 19 Jun 2000, Andreas Beck wrote:
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev.
I am using a cvs-version of ggi and libart (both from last week) to
write to a memory visual and copy that to the screen visual.
O.K. - so you don't do any
"Jon M. Taylor" wrote:
On Mon, 19 Jun 2000, Andreas Beck wrote:
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev.
I am using a cvs-version of ggi and libart (both from last week) to
write to a memory visual and copy that to the screen visual.
O.K. - so you
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Tue, 20 Jun 2000, Jon M. Taylor wrote:
Hmm - this should never flicker. What does flicker - only the changing
parts or the whole screen ?
I think what he is seeing is the fake double-buffer hack I put
into GGIMesa.
Hi Jon!
I don't
GGI itself is targetted at X or /dev/fb. (We do have a GGIMesa Renderer but
that's a totally different story).
BTW: Did the fixed GGIMesa tree I sent you work for you ?
CU, ANdy
--
= Andreas Beck| Email : [EMAIL PROTECTED] =
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev.
I am using a cvs-version of ggi and libart (both from last week) to
write to a memory visual and copy that to the screen visual.
O.K. - so you don't do any screen-switching with Set*Frame - right ?
So we have only
Tobias Hunger writes:
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev. I am
using a cvs-version of ggi and libart (both from last week) to write to a
memory visual and copy that to the screen visual. So we have only one call to
'flush' per frame. The problem:
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
On Mon, 19 Jun 2000, Andrew Apted wrote:
Tobias Hunger writes:
If Berlin can do double buffering, then that might fix the problem.
Berlin is supposed to do doublebuffering... I think there is no option to
even turn it off.
What framebuffer
Andrew Apted wrote:
Tobias Hunger writes:
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev. I am
using a cvs-version of ggi and libart (both from last week) to write to a
memory visual and copy that to the screen visual. So we have only one call to
'flush'
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1
Hello!
I am trying to run berlin (http://www.berlin-consortium.org/) on fbdev. I am
using a cvs-version of ggi and libart (both from last week) to write to a
memory visual and copy that to the screen visual. So we have only one call to
'flush'
15 matches
Mail list logo