Folks, I need some help.
I'm trying to get X-Windows up on my system, and am not having much
luck. Whenever I invoke "startx", the X server starts-up fine, the
screen switches into a graphics mode, and I get a very odd image. The
image consists of a series of vertical bars, generally alternating
between blue and black. There are some other effects, but that is the
most distinguishing (to me, at least). The pointer graphic is clear
and undistorted. When I press the left button on the mouse, I can see
a distorted menu appear on the screen. In higher resolutions, it appears
that the menu is actually being displayed several times.
One person I spoke with suggested that some of the timing information
in the video modes was incorrect. I've tried several different modes,
and each generates nearly identical images, only in different
resolutions. Given that I'm using a multisync monitor, I wouldn't
expect that such an effect would be caused by odd timings. Another
suggested that the monitor was trying to display an image outside the
proper scan rates. I've identified the proper scan rates for my
monitor as identified in the monitor's manual, which is supported by
the XFree86 documentation. So unless the server is ignoring the sync
rates I've identified, that shouldn't be a problem.
Best as I can tell, I'm using Slackware 2.2.0 (well, that's the version
on Setup and PkgTool), which provided v1.2.1 of the Linux kernel. I
have recompiled the kernel since the initial installation. I appear to
have XFree86 v3.1.1 (see below) and am using the XF86_P9000 server. I
used the "xf86config" utility provided with the Slackware distribution
to generate the base "XF86Config" file, to which I then made a few
changes (described below).
Below are what I believe to be the relevant portions of /etc/XF86Config
(prefixed by "C>" to clarify):
C>Section "Monitor"
C> Identifier "CTX CMS-1561"
C> VendorName "CTX"
C> ModelName "CMS-1561"
C> HorizSync 30-60
C> VertRefresh 50-90
All video mode definitions remained unchanged (none commented out, etc.).
C>EndSection
Generic VGA device definition omitted here, since I don't think it's used.
C>Section "Device"
C> Identifier "Diamond Viper VLB 1MB"
C> VendorName "Diamond"
C> BoardName "Viper VLB 1MB"
C> VideoRam 1024
C> Ramdac "bt485"
Added the RAMDAC definition after taking my system apart and inspecting
the card itself. According to the documentation, a RAMDAC definition
isn't necessary. Is it, in this case? I've actually tried it both
ways, and it doesn't appear to make a difference.
I tried adding the following line to see if it made a difference.
C> Option "vram_128"
It didn't, so I removed it.
C> Chipset "vipervlb"
C>EndSection
I'm using the XF86_P9000 server, so I assume that the only screen
section definition that actually takes effect is the following:
C>Section "Screen"
C> Driver "accel"
C> Device "Diamond Viper VLB 1MB"
C> Monitor "CTX CMS-1561"
C> Subsection "Display"
C> Depth 8
C> Modes "640x480" "800x600" "1024x768"
C> ViewPort 0 0
C> Virtual 1024 768
C> EndSubsection
C> Subsection "Display"
C> Depth 16
C> Modes "640x480" "800x600"
C> ViewPort 0 0
C> Virtual 800 600
C> EndSubsection
C> Subsection "Display"
C> Depth 32
C> Modes "640x400"
C> ViewPort 0 0
C> Virtual 640 400
C> EndSubsection
C>EndSection
"/usr/X11/bin/X -probeonly" gives me the following (prefixed with "P>"
to clarify):
P>XFree86 Version 3.1.1 / X Window System Something I find interesting: when I use SuperProbe to identify my S>SuperProbe Version 2.3 Now, I really don't know much about hardware, but the above appears to Many thanks in advance to any and all that might offer assistance. -Steve Coile
P>(protocol Version 11, revision 0, vendor release 6000)
P>Operating System: Linux
P>Configured drivers:
P> P9000: accelerated server for Weitek P9000 graphics adaptors (Patchlevel 0)
P> Supported vendors (specify on chipset line):
P> viperpci Diamond Viper PCI
P> vipervlb Diamond Viper VLB
P> orchid_p9000 Orchid P9000
P>(using VT number 7)
P>
P>XF86Config: /etc/XF86Config
P>(**) stands for supplied, (--) stands for probed/default values
P>(**) Mouse: type: Microsoft, device: /dev/mouse, baudrate: 1200,
P> 3 button emulation
P>(**) P9000: Graphics device ID: "Diamond Viper VLB 1MB"
P>(**) P9000: Monitor ID: "CTX CMS-1561"
P>(--) P9000: Mode "1024x768" needs hsync freq of 62.50 kHz. Deleted.
P>(--) P9000: Mode "1280x1024" needs hsync freq of 64.25 kHz. Deleted.
P>(--) P9000: Mode "1280x1024" needs hsync freq of 78.86 kHz. Deleted.
P>(**) FontPath set to "/usr/X11R6/lib/X11/fonts/misc/,/usr/X11R6/lib/X11/fonts/Type1/,/usr/X11R6/
P>(**) P9000: Vendor/chipset is vipervlb (Diamond Viper VLB)
P>P9000: MemBase not specified. Using 0x80000000 as a default.
P>(**) P9000: (mem: 1024k numclocks: 7 vendor: vipervlb membase: 0x80000000)
P>(--) P9000: Maximum allowed dot-clock: 135.000 MHz
P>(**) P9000: Mode "640x480": mode clock = 31.500
P>(**) P9000: Mode "800x600": mode clock = 50.000
P>(--) P9000: The dimensions of mode 800x600 do not match those of
P> the first valid mode (640x480).
P>(--) P9000: Removing mode "800x600" from list of valid modes.
P>(**) P9000: Mode "1024x768": mode clock = 75.000
P>(--) P9000: The dimensions of mode 1024x768 do not match those of
P> the first valid mode (640x480).
P>(--) P9000: Removing mode "1024x768" from list of valid modes.
P>(**) P9000: Virtual resolution set to 640x480
P>(--) P9000: Using hardware cursor
P>(--) P9000: Using 8 bits per RGB value
video hardware, I get the following (prefixed by "S>" to clarify):
[Copyright, etc. removed for brevity. -sjc]
S>
S>First video: Super-VGA
S> Chipset: Tseng ET4000
S> Memory: 256 Kbytes
S> RAMDAC: Generic 8-bit pseudo-color DAC
S> (with 6-bit wide lookup tables (or in 6-bit mode))
S> Attached graphics coprocessor:
S> Chipset: Chips & Technologies 82C480 (8514/A clone)
S> Memory: 512 Kbytes
be blindingly wrong. I actually physically examined my card and
identified the Weitek POWER 9000 chip. I saw nothing that even
remotely resembled "Tseng ET4000". I also identified the Bt485KPJ135
RAMDAC; is that considered a "generic" RAMDAC??? I have no idea what
the "attached graphics coprocessor" refers to. There *was* another
large chip (i.e. one that struck me as being significant) on the card,
but it also said "Weitek" on it ("Weitek 5186-080-PFP 1641F0016" to be
exact). Does SuperProbe's report have any basis?