Update README source for new release notes.

1 files changed, 49 insertions, 35 deletions

diff --git a/README.sgml b/README.sgml
index fd52608c..4a67017f 100644
--- a/README.sgml
+++ b/README.sgml
@@ -5,14 +5,14 @@
 <article>
 <title>Information for Intel graphics driver users
 <author>Eric Anholt
-<date>2006-08-04
+<date>2006-04-02
 <toc>
 
 <sect>Introduction
 <p>
 This document provides a brief summary of the Intel graphics support provided
 by the xf86-video-intel driver.  More information can also be found in the
-<htmlurl name="i810(4)" url="i810.4.html"> manual page.
+<htmlurl name="intel(4)" url="intel.4.html"> manual page.
 
 <sect>Supported Hardware
 <p>
@@ -41,8 +41,7 @@ by the xf86-video-intel driver.  More information can also be found in the
   <item>Hardware cursor support to reduce sprite flicker.
   <item>Textured video XV implementation on i915 through i965.
   <item>Hardware overlay XV implementation up through i945.
-  <item>Screen resize and rotation on chipsets up through i945.
-  <item>Screen resize on i965.
+  <item>Screen resize and rotation.
 </itemize>
 
 
@@ -53,6 +52,13 @@ by the xf86-video-intel driver.  More information can also be found in the
   <item>This driver  requires kernel support for AGP, which is
 	included in Linux kernels 2.3.42 and higher, and FreeBSD 4.1
 	and higher.
+  <item>This driver may be built against xserver 1.2.0 or newer.  If built
+  from git source against an older server, a copy of xserver 1.3.0 source is
+  needed to complete the build, which is chosen with the --with-xserver-source
+  argument to ./configure.  The compatibility mode with xserver 1.2.0 will have
+  limited functionality, in particular by not supporting RandR 1.2 or the
+  new Damage 1.1 updated needed for 3D applications to work with display
+  rotation.
 </itemize>
 
 
@@ -63,50 +69,58 @@ initialize the card.  The only lines you should need in the "Device"
 section of your xorg.conf file are:
 <verb>
        Section "Device"
-           Identifier "Intel i810"
-           Driver     "i810"
+           Identifier "intel"
+           Driver     "intel"
        EndSection
 </verb>
 
-In order to use most resolutions, it is necessary to install the
-"agpgart.o" module.  You will probably have to compile the module yourself
-(see the notes in the module).
-
 <sect>Driver Options
 <p>
-Please refer to the <htmlurl name="i810(4)" url="i810.4.html"> manual page
+Please refer to the <htmlurl name="intel(4)" url="intel.4.html"> manual page
 for information on configuration options.
 
 
 <sect>Known Limitations
 <p>
 <itemize>
-  <item>Many systems with Intel graphics have issues with setting video modes
-at larger than some small maximum resolution.  This is not fixed in the current
-release, but is being actively worked on in the modesetting branch.
-  <item>Bug #5795: Some systems have issues with VT switching.  This should
-be fixed with the modesetting brach integration.
-  <item>Bug #5817: Hotkey switching from LVDS to CRT breaks CRT output.  This
-is a known issue, but will not be fixed in the current release.
-  <item>Bug #6635: Video is output from an incorrect offset in the framebuffer.
-This is expected to be fixed with the modesetting branch integration.
-  <item>GL_EXT_texture_compression_s3tc is not supported.  We can't support the
-extension due to patent restrictions on compression, but may be able to support
-an option for partial extension support in the future.  For now, this prevents
-Quake4 and some other games from running.
-  <item>Some X Test Suite cases sometimes fail due to a timeout. These cases
-include: Xt8/XtResizeWindow, Xt8/XtQueryGeometry, Xt9/XtAppAddInput,
-Xt9/XtRemoveInput, Xt9/XtAppAddTimeOut, Xt9/XtRemoveTimeOut, Xt9/XtAddGrab,
-Xt9/XtRemoveGrab.
-  <item>Some X Test Suite cases fail in 64-bit mode: Xlib9/XDrawArc,
-XDrawImageString, XDrawLine, XDrawRectangle, XDrawSegments, XFillArc,
-XFillPolygon, XFillRectangle, XPutImage, Xt11/XtVaGetSubresources,
-XtSetSubvalues, and XtVaSetSubvalues.
-  <item>Some GLEAN test cases fail if DRI is enabled: pointAtten,
-readPixSanity, texCombine, texCube, texEnv, texgen, coloredTexPerf2, and
-coloredLitPerf2.
+  <item>Bug #8534: i830 laptop panels not supported.  The driver
+  will fail to recognize them, and only function through VGA output.  Two
+  partial DVO chip drivers (ivch and ch7017) are included which contain some of
+  the code necessary for i830 laptop panel support, but some I2C debugging will
+  be necessary to get those drivers to attach.
+  <item>No support for "zaphod mode" dualhead.  This is the mode in which two
+  Device sections are placed in the config file, and doesn't support DRI or
+  many other features.  Instead, only "MergedFB-style" dualhead is supported.
+  <item>No support for X Screens larger than 2048 pixels in either direction
+  before the 965.  This reflects hardware limitations in the x direction on
+  those older chips, and limits dualhead functionality.  It may be possible to
+  extend the limit vertically on these older chips.
+  <item>i855 XV may cause hangs.  This was present in the previous release,
+  and no workaround is known.
+  <item>SDVO TV-out cards not supported.  This should be fixed in the next
+  release.
+  <item>Gray output with integrated TV-out.
+  <item>EXA support unstable on i845.
 </itemize>
 
+<sect>Common issues not caused by the driver
+<itemize>
+  <item>Font sizes (DPI) are wrong.  Some displays incorrectly report their
+  physical size, which is harmless on most OSes that always assume 96dpi
+  displays.  This can be fixed through quirks for specific monitors in the X
+  Server, and the output of xrandr --prop along with a physical measurement of
+  the screen size in a bug report against the server can help get that fixed.
+  <item>gnome-panel is located in the middle of the screen.  gnome-panel places
+  itself within head #0's boundaries, which doesn't work well with a second
+  head covering the same area as head #0 but larger.
+  <item>Older resolution-changing applicationss have poor results in
+  multihead systems.  Previous extensions such as RandR 1.1 exposed only a
+  single output to client programs, and those requests map poorly to multi-head
+  systems.  Currently, those requests map to just one of the outputs in the
+  RandR 1.2 environment, and those applications need to be updated to RandR 1.2
+  API when available for better results.
+  <item>
+</itemize>
 
 <sect>Author
 <p>