Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Signed-off-by: Alan Coopersmith <alan.coopersmith@oracle.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

While commit 11b230ba removed most usage of AC_CHECK_FILE, one more of
it was left in the compilation build script. This will hopefully fix
the cross build failure for good.
Signed-off-by: Helmut Grohne <helmut@subdivi.de>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Move acquiring of the option settings into r128AcquireOption and
r128UMSOption functions.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Performed DRI1 and EXA / XAA initialization code refactoring.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

The turn on timing is premature. Wait until completion of the
initialization process.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

The previous implementation had duplicated code depending on which
standard is being used (i.e., DRI1 and / or EXA).
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Performed code refactoring and made tweaks to initialization behavior.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Moving them away from R128ScreenInit callback function. This is how
other DDXs do it.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

It is no longer used.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Since SWCursor option is referenced only once, it should be handled
inside R128PreInit callback function rather than R128ScreenInit
callback function.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Since AccelMethod and RenderAccel options are referenced only once,
it should be handled inside R128PreInit callback function rather
than R128ScreenInit callback function.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Since NoAccel option is referenced only once, it should be handled
inside R128PreInit callback function rather than R128ScreenInit
callback function.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Starting the process of separating legacy XAA and newer EXA
initialization code.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Added a temporary fix for some RAGE 128 Pro models misidentifying
VGA as DVI and not able to automatically select the proper screen
resolution via the use of EDID (Extended Display Identification).
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Stop the use of #include to include r128_exa_render.c inside
r128_exa.c.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Dylan Aïssi <bob.dybian@gmail.com>

configure.ac fails to cross build from source, because it abuses
AC_CHECK_FILE for finding headers. AC_CHECK_FILE is supposed to find
files on the host system. You cannot usually expect the host system
to have headers, so this use is incorrect. It happens to also break
cross compilation.
Signed-off-by: Helmut Grohne <helmut@subdivi.de>

Due to faulty assumptions made by commit 2072d9aa, many RAGE 128 Pro
VGA display will be detected as DVI, and causes out of range display
error with the attached monitor. While this commit is less than
perfect and only works on x86 platform, it avoids causing out of
range display error. With this fix, there is no longer the need to
specify the supported frequency range of the monitor inside xorg.conf.
The code was tested on RAGE 128 Pro Ultra 16 MB AGP.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

The variable sizes were not matching the intended print format.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Deprecating R128TRACE macro in favor of DEBUG macro used inside
OpenChrome DDX.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Merge it into R128LegacyMS function.
Signed-off-by: Kevin Brace <kevinbrace@gmx.com>

Apparently this is needed in src/atipcirename.h since xorg-server
1.20 in order to still build this driver.
Signed-off-by: Stefan Dirsch <sndirsch@suse.de>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>