Linux - X configuration options

Answer ID 174
Published 11/30/2004 09:35 AM
Updated 02/25/2008 10:48 AM

X configuration options


The following driver options are supported by the NVIDIA X driver.
They may be specified either in the Screen or Device sections of the X
config file.

Option "NvAGP" "integer"
   Configure AGP support. Integer argument can be one of:
   0 : disable agp
   1 : use NVIDIA's internal AGP support, if possible
   2 : use AGPGART, if possible
   3 : use any agp support (try AGPGART, then NVIDIA's AGP)
   Please note that NVIDIA's internal AGP support cannot
   work if AGPGART is either statically compiled into your
   kernel or is built as a module, but loaded into your
   kernel (some distributions load AGPGART into the kernel
   at boot up).  Default: 3 (the default was 1 until after
   1.0-1251).

Option "NoLogo" "boolean"
   Disable drawing of the NVIDIA logo splash screen at
   X startup.  Default: the logo is drawn.

Option "RenderAccel" "boolean"
   Enable or disable hardware acceleration of the RENDER
   extension.  THIS OPTION IS EXPERIMENTAL.  ENABLE IT AT YOUR
   OWN RISK.  There is no correctness test suite for the
   RENDER extension so NVIDIA can not verify that RENDER
   acceleration works correctly.   Default: hardware
   acceleration of the RENDER extension is disabled.

Option "NoRenderExtension" "boolean"
   Disable the RENDER extension.  Other than recompiling
   the X-server, XFree86 does not seem to have another way of
   disabling this.  Fortunatly, we can control this from the
   driver so we export this option.  This is useful in depth
   8 where RENDER would normally steal most of the default
   colormap. Default: RENDER is offered when possible.

Option "UBB" "boolean"
   Enable or disable Unified Back Buffer on any Quadro
   based GPUs (Quadro4 NVS excluded); please see
   Appendix M for a description of UBB.  This option has
   no affect on non-Quadro chipsets.  Default: UBB is on
   for Quadro chipsets.

Option "NoFlip" "boolean"
   Disable OpenGL flipping; please see Appendix M for
   a description.  Default: OpenGL will swap by flipping
   when possible.

Option "DigitalVibrance" "integer"
   Enables Digital Vibrance Control.  The range of valid
   values are 0 through 255.  This feature is not available
   on products older than GeForce2.  Default: 0.

Option "Dac8Bit" "boolean"
   Most Quadro parts by default use a 10 bit color look
   up table (LUT) by default; setting this option to TRUE forces
   these graphics chips to use an 8 bit (LUT).  Default:
   a 10 bit LUT is used, when available.

Option "Overlay" "boolean"
   Enables RGB workstation overlay visuals.  This is only
   supported on Quadro4 and Quadro FX chips (Quadro4 NVS
   excluded) in depth 24.  This option causes the server to
   advertise the SERVER_OVERLAY_VISUALS root window property
   and GLX will report single and double buffered, Z-buffered
   16 bit overlay visuals.  The transparency key is pixel
   0x0000 (hex).  There is no gamma correction support in
   the overlay plane.  This feature requires XFree86 version
   4.1.0 or newer (or the Xorg X server).  NV17/18 based
   Quadros (ie. 500/550 XGL) have additional restrictions,
   namely, overlays are not supported in TwinView mode
   or with virtual desktops larger than 2046x2047 in any
   dimension (eg.  it will not work in 2048x1536 modes).
   Quadro 7xx/9xx and Quadro FX will offer overlay visuals
   in these modes (TwinView, or virtual desktops larger
   than 2046x2047), but the overlay will be emulated with
   a substantial performance penalty.  RGB workstation
   overlays are not supported when the Composite extension is
   enabled.  Default: off.

Option "CIOverlay" "boolean"
   Enables Color Index workstation overlay visuals with
   identical restrictions to Option "Overlay" above.
   The server will offer visuals both with and without a
   transparency key.  These are depth 8 PseudoColor visuals.
   Enabling Color Index overlays on X servers older than
   XFree86 4.3 will force the RENDER extension to be disabled
   due to bugs in the RENDER extension in older X servers.
   Color Index workstation overlays are not supported when the
   Composite extension is enabled.  Default: off.

Option "TransparentIndex" "integer"
   When color index overlays are enabled, use this option
   to choose which pixel is used for the transparent pixel
   in visuals featuring transparent pixels.  This value
   is clamped between 0 and 255 (Note: some applications
   such as Alias's Maya require this to be zero
   in order to work correctly).  Default: 0.

Option "OverlayDefaultVisual" "boolean"
   When overlays are used, this option sets the default
   visual to an overlay visual thereby putting the root
   window in the overlay.  This option is not recommended
   for RGB overlays.  Default: off.

Option "SWCursor" "boolean"
   Enable or disable software rendering of the X cursor.
   Default: off.

Option "HWCursor" "boolean"
   Enable or disable hardware rendering of the X cursor.
   Default: on.

Option "CursorShadow" "boolean" Enable or disable use of a
   shadow with the hardware accelerated cursor; this is a
   black translucent replica of your cursor shape at a
   given offset from the real cursor.  This option is
   only available on GeForce2 or better hardware (ie
   everything but TNT/TNT2, GeForce 256, GeForce DDR and
   Quadro).  Default: no cursor shadow.

Option "CursorShadowAlpha" "integer"
   The alpha value to use for the cursor shadow; only
   applicable if CursorShadow is enabled.  This value must
   be in the range [0, 255] -- 0 is completely transparent;
   255 is completely opaque.  Default: 64.

Option "CursorShadowXOffset" "integer"
   The offset, in pixels, that the shadow image will be
   shifted to the right from the real cursor image; only
   applicable if CursorShadow is enabled.  This value must
   be in the range [0, 32].  Default: 4.

Option "CursorShadowYOffset" "integer"
   The offset, in pixels, that the shadow image will be
   shifted down from the real cursor image; only applicable
   if CursorShadow is enabled.  This value must be in the
   range [0, 32].  Default: 2.

Option "ConnectedMonitor" "string"
   Allows you to override what the NVIDIA kernel module
   detects is connected to your video card.  This may
   be useful, for example, if you use a KVM (keyboard,
   video, mouse) switch and you are switched away when
   X is started.  In such a situation, the NVIDIA kernel
   module cannot detect what display devices are connected,
   and the NVIDIA X driver assumes you have a single CRT.

   Valid values for this option are "CRT" (cathode ray
   tube), "DFP" (digital flat panel), or "TV" (television);
   if using TwinView, this option may be a comma-separated
   list of display devices; e.g.: "CRT, CRT" or "CRT, DFP".

   NOTE: anything attached to a 15 pin VGA connector is
   regarded by the driver as a CRT.  "DFP" should only be
   used to refer to flatpanels connected via a DVI port.

   Default: string is NULL.

Option "UseEdidFreqs" "boolean"
   This option causes the X server to use the HorizSync
   and VertRefresh ranges given in a display device's EDID,
   if any.  EDID provided range information will override
   the HorizSync and VertRefresh ranges specified in the
   Monitor section.  If a display device does not provide an
   EDID, or the EDID does not specify an hsync or vrefresh
   range, then the X server will default to the HorizSync
   and VertRefresh ranges specified in the Monitor section.

Option "IgnoreEDID" "boolean"
   Disable probing of EDID (Extended Display Identification
   Data) from your monitor.  Requested modes are compared
   against values gotten from your monitor EDIDs (if any)
   during mode validation.  Some monitors are known to lie
   about their own capabilities.  Ignoring the values that
   the monitor gives may help get a certain mode validated.
   On the other hand, this may be dangerous if you do not
   know what you are doing.  Default: Use EDIDs.

Option "NoDDC" "boolean"
   Synonym for "IgnoreEDID"

Option "FlatPanelProperties" "string"
   Requests particular properties of any connected flat
   panels as a comma-separated list of property=value pairs.
   Currently, the only two available properties are 'Scaling'
   and 'Dithering'.   The possible values for 'Scaling' are:
   'default' (the driver will use whatever scaling state
   is current), 'native' (the driver will use the flat
   panel's scaler, if it has one), 'scaled' (the driver
   will use the NVIDIA scaler, if possible), 'centered'
   (the driver will center the image, if possible),
   and 'aspect-scaled' (the driver will scale with the
   NVIDIA scaler, but keep the aspect ratio correct).
   The possible values for 'Dithering' are: 'default'
   (the driver will decide when to dither), 'enabled' (the
   driver will always dither when possible), and 'disabled'
   (the driver will never dither).  If any property is not
   specified, it's value shall be 'default'.  An example
   properties string might look like:

   "Scaling = centered, Dithering = enabled"

Option "UseInt10Module" "boolean"
   Enable use of the X Int10 module to soft-boot all
   secondary cards, rather than POSTing the cards through
   the NVIDIA kernel module.  Default: off (POSTing is done
   through the NVIDIA kernel module).

NOTE: For details on Twinview, see the article titled "Linux - Configuring Twinview."

Option "TwinView" "boolean"
   Enable or disable TwinView. Default: TwinView is disabled.

Option "TwinViewOrientation" "string"
   Controls the relationship between the two display devices
   when using TwinView.  Takes one of the following values:
   "RightOf" "LeftOf" "Above" "Below" "Clone". Default: string is NULL.

Option "SecondMonitorHorizSync" "range(s)"
   This option is like the HorizSync entry in the Monitor
   section, but is for the second monitor when using
   TwinView. Default:
   none.

Option "SecondMonitorVertRefresh" "range(s)"
   This option is like the VertRefresh entry in the Monitor
   section, but is for the second monitor when using
   TwinView. Default:
   none.

Option "MetaModes" "string"
   This option describes the combination of modes to use
   on each monitor when using TwinView. Default: string is NULL.

Option "NoTwinViewXineramaInfo" "boolean"
   When in TwinView, the NVIDIA X driver normally provides
   a Xinerama extension that X clients (such as window
   managers) can use to to discover the current TwinView
   configuration.  Some window mangers can get confused by
   this information, so this option is provided to disable
   this behavior.  Default: TwinView Xinerama information
   is provided.

Option "TVStandard" "string"
   Please see the article titled "Linux - Configuring TV-Out."

Option "TVOutFormat" "string"
  Please see the article titled "Linux - Configuring TV-Out."

Option "TVOverScan" "Decimal value in the range 0.0 to 1.0"
   Valid values are in the range 0.0 through 1.0. Please see the article titled "Linux - Configuring TV-Out."

Option "Stereo" "integer"
   Enable offering of quad-buffered stereo visuals on Quadro.
   Integer indicates the type of stereo glasses being used:
 
   1 - DDC glasses.  The sync signal is sent to the glasses
  via the DDC signal to the monitor.  These usually
  involve a passthrough cable between the monitor and
  video card.

   2 - "Blueline" glasses.  These usually involve
  a passthrough cable between the monitor and video
  card.  The glasses know which eye to display based
  on the length of a blue line visible at the bottom
  of the screen.  When in this mode, the root window
  dimensions are one pixel shorter in the Y dimension
  than requested.  This mode does not work with virtual
  root window sizes larger than the visible root window
  size (desktop panning).

   3 - Onboard stereo support.  This is usually only found
  on professional cards.  The glasses connect via a
  DIN connector on the back of the video card.

   4 - TwinView clone mode stereo (aka "passive" stereo).
  On video cards that support TwinView, the left eye
  is displayed on the first display, and the right
  eye is displayed on the second display.  This is
  normally used in conjuction with special projectors
  to produce 2 polarized images which are then viewed
  with polarized glasses.  To use this stereo mode,
  you must also configure TwinView in clone mode with
  the same resolution, panning offset, and panning
  domains on each display.

   Stereo is only available on Quadro cards.  Stereo
   options 1, 2, and 3 (aka "active" stereo) may be used
   with TwinView if all modes within each metamode have
   identical timing values.  Please see the article titled "Linux - Programming modes"

for suggestions on making sure the
   modes within your metamodes are identical.  The identical
   modeline requirement is not necessary for Stereo option 4
   ("passive" stereo).  Currently, stereo operation may
   be "quirky" on the original Quadro (NV10) chip and
   left-right flipping may be erratic.  We are trying
   to resolve this issue for a future release.  Default:
   Stereo is not enabled.

   Stereo options 1, 2, and 3 (aka "active" stereo) are not
   supported on Digital Flatpanels.

Option "AllowDFPStereo" "boolean"
   By default, the NVIDIA X driver performs a check which
   disables active stereo (stereo options 1, 2, and 3)
   if the X screen is driving a DFP.  The "AllowDFPStereo"
   option bypasses this check.

Option "NoBandWidthTest" "boolean"
   As part of mode validation, the X driver tests if a
   given mode fits within the hardware's memory bandwidth
   constraints.  This option disables this test.  Default:
   the memory bandwidth test is performed.

Option "IgnoreDisplayDevices" "string"
   This option tells the NVIDIA kernel module to completely
   ignore the indicated classes of display devices when
   checking what display devices are connected.  You may
   specify a comma-separated list containing any of "CRT",
   "DFP", and "TV".

   For example:

  Option "IgnoreDisplayDevices" "DFP, TV"

   will cause the NVIDIA driver to not attempt to detect
   if any flatpanels or TVs are connected.

   This option is not normally necessary; however, some video
   BIOSes contain incorrect information about what display
   devices may be connected, or what i2c port should be
   used for detection.  These errors can cause long delays
   in starting X.  If you are experiencing such delays, you
   may be able to avoid this by telling the NVIDIA driver to
   ignore display devices which you know are not connected.

   NOTE: anything attached to a 15 pin VGA connector is
   regarded by the driver as a CRT.  "DFP" should only be
   used to refer to flatpanels connected via a DVI port.

Option "MultisampleCompatibility" "boolean"
   Enable or disable the use of separate front and back
   multisample buffers.  This will consume more memory
   but is necessary for correct output when rendering to
   both the front and back buffers of a multisample or
   FSAA drawable.  This option is necessary for correct
   operation of SoftImage XSI.  Default: a singlemultisample
   buffer is shared between the front and back buffers.

Option "NoPowerConnectorCheck" "boolean"
   The NVIDIA X driver will abort X server initialization
   if it detects that a GPU that requires an external power
   connector does not have an external power connector
   plugged in.  This option can be used to bypass this test.
   Default: the power connector test is performed.

Option "XvmcUsesTextures" "boolean"
   Forces XvMC to use the 3D engine for XvMCPutSurface
   requests rather than the video overlay.  Default: video
   overlay is used when available.

Option "AllowGLXWithComposite" "boolean"
   Enables GLX even when the Composite X extension is loaded.
   ENABLE AT YOUR OWN RISK.  OpenGL applications will not
   display correctly in many circumstances with this setting
   enabled.  Default: GLX is disabled when Composite is
   loaded.

Was this answer helpful?
Your rating has been submitted, please tell us how we can make this answer more useful.

LIVE CHAT

Chat online with one of our support agents

CHAT NOW

ASK US A QUESTION

Contact Support for assistance

CONTACT US