xf86-video-ati: Changes to 'refs/tags/xf86-video-ati-6_5_8_1'

Dave Airlie airlied at kemper.freedesktop.org
Tue Sep 12 06:50:37 EEST 2006

Tag 'xf86-video-ati-6_5_8_1' created by Dave Airlie <airlied at starflyer-ubuntu.(none)> at 2006-09-11 18:50 -0700

ATI for X.org 7.0 -

Changes since the dawn of time:
Adam Jackson:
      Bugzilla #849: Fix DRI pagesize assumptions in radeon and r128 drivers.
      Bug #848: Clean up gatos notice messages.
      Bug #400 (partial): Driver fixes for the dlloader. When using dlloader, all
      Change several LoaderSymbol calls introduced by the bug #400 patch to
      Bug #978: Silence unresolved symbol warnings from the ati driver. Reported
      Bug #978: Silence more unresolved symbol warnings from the ati driver.
      Factor out R{128,ADEON}Chipsets into their own headers so they can be
      oops, these didn't added somehow.
      Bug #1061 (documentation-only update): Document the RenderAccel option in
      Back out my last change, dupe.
      Bug #1192: Remove cfb support from drivers where its use is an option.
      Convert the ATI drivers to ANSI function prototypes.
      Bug #1252, #1253, #1255, #1256: Various typo fixes from Dave Jones.
      Correct radeon man page, IGP is supported for 3D but R400 is not.
      Bug #2008: Fix gcc4 compile problem.
      Bug #2567: Fix reversed memset arguments. (David Krause)
      Bug #3163: Create new DRIINFO_*_VERSION macros to indicate the version
      Typo fix from bug #3163 commit
      Bug #3626: _X_EXPORT tags for video and input drivers.
      Bug #3628: Fix video driver submodule visibility.
      Prep for modular builds by adding guarded #include "config.h" everywhere.
      Post-merge fixups:
      Explicit license statement.
      Bug #1109: Fix VGA init on Rage Mobility 7500 (Marc Le France)
      Bug #4148: Bump required DRM version for r300 to 1.17.
      Disable the {Open,Close}FullScreen DRI protocol. Remove empty FullScreen
      Bug #4001: Don't crash when PanelSize is unset. (Alan Coopersmith)
      More 1.7 braindamage: define EXTRA_DIST in terms of @DRIVER_NAME@ instead
      Bug #4829: Remove some useless DRI symbol checks.
      Only build dlloader modules by default.
      Bump libdrm dep to 2.0.
      Bug #1106: Fix memory allocation for Render. (Michel Daenzer)
      Bug #3869: Don\'t walk off the end of an array in connector detection.
      Bug #1760: Fix ati wrapper multihead to be dlloader friendly.
      Stub COPYING files
      Bug #5523: Make radeon compile without USE_XAA. (Tilman Sauerbeck)

Alan Coopersmith:
      Sync with 6.8.2 branch: Bugzilla #1306
      Sync with 6.8.2 branch: Bug #591
      #include "extensions/dpms.h" -> <X11/extensions/dpms.h>
      Only include radeon_dri.h & radeon_sarea.h if XF86DRI is defined.
      Use sed & cpp to substitute variables in driver man pages
      Use @DRIVER_MAN_SUFFIX@ instead of $(DRIVER_MAN_SUFFIX) in macro
      Sun bug #6325220: Xorg dumps core in dual monitor configuration with radeon
      Change *mandir targets to use new *_MAN_DIR variables set by xorg-macros.m4
      Bugzilla #5238 <https://bugs.freedesktop.org/show_bug.cgi?id=5238> Patch
      Typo fixes, mailing list & url updates, and other changes to prepare for
      When failing to parse MetaModes, reset CRT2pScrn to NULL after free'ing it

Alan Hourihane:
      Egbert's 64bit fixes for mixed 32/64bit clients

Alex Deucher:
      - Add Radeon DynamicClocks option
      - add gamma correction support to the radeon video overlay
      Update comment
      - fix resume on radeons (bug 1220)
      - Add Thomas' latest sis MergedFB enhancements to the radeon driver
      - Fix DFP detection in r128 driver (bug 1386, Klaus Kusche)
      - make the overlay work on crtc2 again
      - only mess with Dynamicclocks on mobility chips (Benh, me) bug 1912
      - add AGP 8x support (Hui Yu, Michel Daenzer) bug 1819
      - missing diff from agp 8x support
      - Add VGAAccess option to r128 driver. Based on Benh's radeon patch. r128
      - fix warning and potential error in radeon_driver.c
      - fix warnings (Ronny V. Vindenes)
      - Add DRI resume support (Micha Feigin) bug 2063
      - Make dualhead ram allocation cleaner and related messages clearer
      - Add support for Dualhead on M3/M4 (bug 1760)
      - Fix MCLK / SCLK calculation (Benjamin Herrenschmidt) bug 2144
      - radeon PLL access errata workarounds (Benjamin Herrenschmidt, acked by
      - Add new Radeon pci ids (ATI devrel), bug 2827
      Sync up radeon mergedfb with Thomas' latest sis mergedfb changes:
      - fix some OUTREGs that should be OUTPLLs in dynamicclocks code
      - Increase the priority of display 0/1 in the memory controller for r3/4xx
      - add support for colortiling on r3/4xx hardware (disabled by default). In
      - small mergedfb fix from Thomas Winischhofer
      - change back part of a comment I seem to have accidently changed.
      - Fix Support for Philips FM1236/F tuner on ATI AIW 9600 XT (Jeff Smith)
      - fix depth tiling for r3/4xx chips (Aapo Tahkola)
      - missing line in mergedfb option parsing. Spotted by Christoph Bauer
      - radeon XPRESS chips are actually RV3xx based
      - fix pitch increment (Aapo Tahkola, Roland Scheidegger)

Alex Williamson:
      Bug #6872: Disable digital output while reading ROM with Radeon 7000.

Benjamin Herrenschmidt:
      Fix EXA acceleration of textures with byteswap or tiling From John Clemens
      The engine setup would trigger a bogus line write before the proper
      Memory map and misc fixes. This completely reworks the way the driver
      Fix page flipping with XAA. The new code ended up calling ShadowFBInit()
      Workaround for rv280 bug which needs the framebuffer memory mapping to be
      Extend the alignement workaround to post-rv280 chips as well (thanks Hui)
      Fix the fix ... I didn't interpret PciInfo->size properly and forgot that
      Bugzilla #5766 <https://bugs.freedesktop.org/show_bug.cgi?id=5766> Patch
      More memory map fixes !!! This time, we add a way to get the DRM kernel
      Use "FBTexPercent" option for XAA too (only for increasing memory reserved
      Add #ifdef XF86DRI guards around references to info->allowPageFlip and
      Stop using xf86PciInfo.h, instead use a local copy of the PCI IDs we need
      Earliest fix introduced a bug due to another chicken & egg problem in card
      Fix typo in the function setting the color space transformation factors for
      Fix various issues with DRI & server recycle by putting DRICloseSreen in
      Add various workarounds that seem to fix some remaning lockup scenarios I'm
      Fix incorrect CRTC2_OFFSET values when using old-style dual head (not
      Add missing return statement (the driver would not start randomly), add
      Clear the offsets when updating the memory map instead of when enabling the
      Call RADEONRestoreCommonRegisters() before RADEONRestoreCrtc2Registers()
      Fix a regression of my latest fix that broke color tiling on some setups.
      Fix enabling of external VGA output on some cards with dual entity setup

Bogdan Diaconescu:
      Added the RageTheatre200 video demodulator support
      Now the detection of the theatre chip is in other module called
      This is the theatre chip detection module
      Added RT200 registers
      Added functions that access the FIFO on RT200
      Use the theatre_detect to determine the chip ID and then load the
      Addeded support for FI1216ME (PAL) version of the tuner, detection of
      Added new options for the ATI driver: RageTheatreMicrocPath and
      Added fifo_read and fifo_write in tge generic bus definition
      Modified the licens type from GPL to a X/MIT one
      Replaced the variadic macros(gcc) by macros according to C99 standard

Daniel Stone:
      Use fp_horz_stretch instead of fp_vert_stretch for horizontal stretch
      Retry CCE idle commands until the (higher) timeout is reached, and properly
      Disable interrupt handling after the DRM handler has been removed,
      Apply Debian patch to re-POST via VBE if driver thinks it has no video
      Release SDA/SDL I2C lines from an asserted-low state after a DDC probe,
      Fix includes right throughout the Xserver tree:
      Change all misc.h and os.h references to <X11/foo.h>.
      Make DGA, CPIO, non-PCI, and TV-Out support detected per-platform, using
      Update autogen.sh to one that does objdir != srcdir.
      Blacklist Radeon M300 from multimedia I2C code. Handle I2C timeouts more
      Fix bug in Radeon connector table parsing -- parse all entries, not just
      Only use vgahw symbols if we're building vgahw for this architecture. TODO:
      Fix 'ugly' warning in r128. (Branden Robinson, Debian #069)
      Define USE_XAA and USE_EXA for the time being.
      Bug #5656: Set all Radeon XPRESS 200 R[CS]4xx-based chips to have IsIGP =

Dave Airlie:
      bugzilla #2057 (https://bugs.freedesktop.org/show_bug.cgi?id=2057)
      Handle VIP timeouts more gracefully -- impose a hard limit of 10ms on
      Add support for allocating PCI GART table in framebuffer memory. This is
      Move GART offset/size outside XAA
      Fix Radeon EXA for when RENDER isn't defined
      Switch radeon driver to using a card type rather than having a IsPCI flag,
      Secure the Radeon PCIE GART tables in framebuffer memory by not allowing
      Add support for backing up the PCIE GART table on VT switch. This makes
      Bugzilla 6672: fix interrupts on radeon since memmap changes
      prepare stable branch for release, pass distcheck
      radeon: force CP and VIP clocks on some r300 and rv100 chips.
      Make the ATI driver use the PCI BAR size to determine the MMIO register
      remove changelog from stable branch
      Bug #6761: Fix font rendering in non-DRI mode by replacing RB2D_DSTCACHE_*
      radeon: fallback render repeat for XAA, this fixes corruptions seen on ubuntu
      update configure.ac for new release

Donnie Berkholz:
      Bug #4831 <https://bugs.freedesktop.org/show_bug.cgi?id=4831>. Check for

Egbert Eich:
      Importing vendor version xf86-012804-2330 on Thu Jan 29 00:06:33 PST 2004
      Importing vendor version xf86-4_3_99_903 on Wed Feb 26 01:21:00 PST 2004
      readding XFree86's cvs IDs
      Importing vendor version xf86-4_4_0 on Wed Mar 3 04:09:24 PST 2004
      Importing vendor version xf86-4_4_99_1 on Sun Mar 14 00:26:39 PST 2004
      Merging XORG-CURRENT into trunk
      Separated Intel drivers from default DriDrivers to avoid building them on
      Improving DPMS handling on VT swich and server termination/abort: previous
      Set fbdev mode as the default mode on PPC (Olaf Hering).
      Made handling of DevelDrivers for x86-64 more conformant to other
      Fixing one of the infamous 'control reaches end of non void function
      Make Xorg the default server to install.
      Don't exclude 'RADEONInfoPtr info = ...' declaration even when not building
      Explicitely disable DPMS power states during close screen. Fix macro that

Eric Anholt:
      DRI XFree86-4_3_99_12-merge import
      Initial revision
      DRI trunk-20040613 import
      Merge DRI-trunk-20040613 changes in programs/Xserver/hw/xfree86/drivers,
      Bug #748:
      Add file missed in last commit: Bug #748:
      Bugzilla #803,804: Fix build in the !XF86DRI case after DRI and Radeon
      Merge and resolve conflicts from DRI/DRM/Mesa 20040721 imports.
      - Add a new option, BuildDevelDRIDrivers, which controls whether
      - Add two new XAA hooks, SetupForCPUToScreenTexture2 and
      Fix unresolved symbol warnings in r128 driver (Adam Jackson)
      Reenable Radeon Render acceleration now that interactions with 3D clients
      Bug #1157: Fix Radeon line acceleration in the DMA case (Michel Daenzer).
      Bug #1156: Fix the build on PPC by splitting the byte-order-setting code
      Bugzilla #1170: Fix an overcounting of DMA in BEGIN_RING that wasn't
      Add missing detection of the PCI ID for 0x5b62, Radeon X600 (Vladimir
      Add support for EXA to the radeon driver. Building EXA and XAA support is
      Attempt to fix up the modular pieces for EXA support. (untested)
      Remove the hardcoded USE_EXA/USE_XAA defines now that they've been
      Bug #2573: Remove the disabling of automatic cache flushing on non-r300
      - Don't try to upload 0 byte-per-pixel (PICT_a1) data using
      Break EXA ABI while we still can. Add coordinates to the UploadToScreen
      Fix Radeon MMIO Render acceleration with EXA by writing out floating- point
      Bug #1028: Don't (incorrectly) mark the 24/8 z/stencil case as slow, which
      Add .cvsignores for drivers.
      Fix a copy'n'paste-o in the EXA render implementation that resulted in
      Bug #5632: Fix up ati.4 to explain that it is a wrapper driver for the

Felix Kuehling:
      Fix for bugzilla https://bugs.freedesktop.org/show_bug.cgi?id=3414 (patch

Hui YU:
      Support for New radeon chips: R420/M18, R423, RV370/M22, RV380/M24, RS300.

Ian Romanick:
      I have resolved this by making all the "Direct rendering disabled" messages

Kaleb Keithley:
      R6.6 is the Xorg base-line
      Initial revision
      Initial revision
      XFree86 Bring the tree up to date for the Cygwin folks
      XFree86 (RC 1)
      XFree86 (RC 2)
      Import most of XFree86 4.4RC3. This import excludes files which have the

Kevin E Martin:
      Fix Radeon crash on server exit or generation
      Disable Render acceleration in Radeon driver temporarily until it no longer
      Fix install problem on platforms not using xorg.cf/xfree86.cf (Bug #339,
      Fix Radeon build problem with #define DoLoadableServer NO
      Fix context switching on Radeons and build problem when XF86DRI is not
      Fix hangs during startup with Radeon driver (Bug #1007, Sjoerd Simons).
      Fix Radeon compilation without Xinerama (Bug #1179, Simon Schubert).
      Updated release notes for bug #999.
      bugzilla #2840 (https://bugs.freedesktop.org/show_bug.cgi?id=2840)
      Make compiling with DRI support configurable/autodetecable. Autodetect DRI
      Fix check for headers. Use {XORG,DRI}_CFLAGS subst instead of replacing
      Fix filename typo
      Various changes preparing packages for RC0:
      Do the following to make the drivers pass distcheck:
      Update package version number for RC1 release.
      Update pkgcheck depedencies to work with separate build roots.
      Fix DoLoadableServer NO build.
      Update package version number for X11R7 RC2 release.
      Bump version numbers for X11R6.9/X11R7 RC2.
      Remove extraneous AC_MSG_RESULT.
      Bump version numbers for X11R6.9/X11R7 RC3.
      Update package version number for X11R7 RC3 release.
      Change *man_SOURCES ==> *man_PRE to fix autotools warnings.
      Allow hard-coded paths to be configurable.
      Add configure options to allow hard-coded paths to be changed.
      Update package version number for final X11R7 release candidate. Bump
      Update package version for X11R7 release.

Markus Kuhn:
      Encoding of numerous files changed to UTF-8

Matthias Hopf:
      Bug #4998: Add ';' and ',' as valid separators for MetaModes as well.

Matthieu Herrb:
      programs/Xserver/hw/xfree86/drivers/ati/radeon_merged_fb.c Add missing
      programs/Xserver/hw/xfree86/drivers/ati/theatre.h Build fix for
      add missing return value on success in RADEONSetupMemXAA().
      missed one missing return value.

Michel Daenzer:
      Use info->ModeReg.surface_cntl to restore byte swapping for the framebuffer
      Hotfix for Eric Anholt's fix for bug #1156 : Clear the
      When direct rendering is enabled, use hostdata blits to transfer data from
      Bug #2576: Add support for ATI RN50/ES1000. (ATI Technologies Inc.)
      Add support for production version of ATI RN50/ES1000. (ATI Technologies
      bugzilla #2844 (https://bugs.freedesktop.org/show_bug.cgi?id=2844)
      bugzilla #3911 (https://bugs.freedesktop.org/show_bug.cgi?id=3911)
      bugzilla #4951 (https://bugs.freedesktop.org/show_bug.cgi?id=4951)
      bugzilla #4916 (https://bugs.freedesktop.org/show_bug.cgi?id=4916) Fix
      bugzilla #4494 (https://bugs.freedesktop.org/show_bug.cgi?id=4494) Call
      bugzilla #988 (https://bugs.freedesktop.org/show_bug.cgi?id=988) Fix typo
      Add Option "FBTexPercent" to override the amount of video RAM reserved for

Michel Dänzer:
      Bug #6939: Disable RENDER acceleration by default on RN50.

Roland Scheidegger:
      fix issues with mergedfb + pageflip (fix up sarea frame values)
      (Stephane Marchesin, me) Add support for color (framebuffer) tiling.
      change option EnableColorTiling to ColorTiling
      fix error with color tiling and hw scrolling, it looks like rv100 doesn't
      do not enable color tiling on r300 chips and newer, since it doesn't work
      Fix building radeon driver with BuildXF86DRI set to NO, the breakage was
      Fix accidentally removed depth pitch adjustment if displayWidth is not a
      Bug #2604: Disable workaround for horrible flickering when scrolling
      Move the RADEONEnablePageFlip/RADEONChangeSurfaces calls back to the end of
      Add pci ids known to exist (see #4284 for instance). There are still
      Add the newly added (and tons of previously forgotten ones) pci ids to the
      Fix various small cosmetic issues. Change a driver message, get the order
      Use programmable tap coefficients for (hopefully) better video image

Søren Sandmann Pedersen:
      Try adding build system for ATI driver again. Who knows, maybe CVS works
      fuck you cvs
      Install theathre drivers in @moduledir@/multimedia
      Make ati driver distcheck
      - For all drivers that have a <driver>.sgml file, add code in their build
      Check in generated README files

Vladimir Dergachev:
      Add more multimedia registers.
      Add registers to facilitate work with stereo 3d.
      programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c Turns out Radeon R420
      programs/Xserver/hw/xfree86/drivers/ati/radeon_dri.c Let through all chips

Weixing Zhang:
      Bug #6939: Disable 3D acceleration on RN50 by setting invalid ClientDriverName.

More information about the xorg-commit mailing list