xenocara/driver/xf86-video-cirrus
matthieu 7846d614e2 xf86-video-cirrus 1.2.1 + one bug fix from git-current.
Tested on alpine gl5446 on alpha.
2008-09-16 22:02:30 +00:00
..
man xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
src xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
aclocal.m4 xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
ChangeLog xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
config.guess
config.h.in xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
config.sub
configure xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
configure.ac xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
COPYING
depcomp
install-sh
ltmain.sh xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
Makefile.am
Makefile.bsd-wrapper
Makefile.in xf86-video-cirrus 1.2.1 + one bug fix from git-current. 2008-09-16 22:02:30 +00:00
missing
README.multihead

WARNING:
On exit XFree86 restores the old mode but it
does not disable the secondary card. As a result
the second card (which has some randonm CRTC timing)
may produce synch pulses which are wrong for your
monitor.

If your monitor does not tolerate bad timing I would
recommend that you don't try Multi-head yet, or at
least, turn it off *before* leaving X.


3.9Nl notice: There is a problem with releasing
of LBX colormap privates which may cause a seg.
fault when the server exits. I got a patch from
Mark Vojkovich which fixes it.

If you have a CL-GD5480 or CL-GD5446 ver. B you
should be able to try Multi-Head with that
card as a secondary. The primary card should not
have any special feature. Some MainBoards may
not do it right too. I tried it with a 5465 (AGP)
as a primary and 5446B (PCI) secondary on an
ABIT LX6 upgraded to the latest bios.

Version A of the 5446 cannot be pure MMIO (The driver 
requires that it is the primary display adapter).
You can distinguish between the rel A and rel B by
running "scanpci -v". Version 1 does not report a
valid BASE1 address.

The bios upgrade lets me chose if the primary
display adapter is on PCI on AGP (I chose
the AGP card as primary). Warning - this setup
will *NOT* work with XFree86 3.3.2.

The only problem that I found so far is with
memory configuration initialization. My bios
does not initialize the secondary display adapter.
This results in a currupted display.

Here are the options:
1. warm booting from Windows 98 may work. I have
not tried that.

2. Take out the other card and leave the 5446
(or 5480) as the primary card. Start the server
and watch it's output. Look for lines like:
  (--) CIRRUS(0): Memory Config reg 1 is 0x1B
  (--) CIRRUS(0): Memory Config reg 2 is 0x21
Remember these values. Now edit XF86Config
and add to the proper device section the lines:
  Option "MemCFG1" "0x1B"
  Option "MemCFG2" "0x21"
These values are the contents of registers SR0F
and SR17. (notice: the 5480 driver does not use
and will not print MemCFG2).

Now you can reinstall the other card (which
must me set as the the primary display adaptor).

A note about configuring multi-head:
There should be a separate section for each
adapter (device), Monitor and Screen. A "ServerLayout"
section is optional (and recommended). To make it
easier for XFree86 to relate device sections to
actual cards, use a BusId line in each device
section. XFree86 3.9N works properly if the config
file is for Multi-head but it finds only one
display adapter.