[ANNOUNCE] xf86-video-ati - release candidate

Gene Heskett gene.heskett at verizon.net
Mon Dec 22 21:25:08 PST 2008

On Monday 22 December 2008, Dave Airlie wrote:
>This is a pre-release of what will be called xf86-video-ati 6.9.1
>Pre-release notes:
>This contains a lot of bugfixes/enhancements, highlights include
>Output support for all radeon chips from r100 to rv730 (mergedfb accel for
> all r600 and newer) - rv710/730 DCE3.2 support
>	- Initial eMac support
>	- better HDMI support
>Enhanced textured video support
>	- Bicubic shaders
>	- Large triangle rendering to decrease tearing
>	- vsynced Xv
>Interlaced mode support
>Lots of EXA correctness and stability fixes
>Big-endian atombios support.
>Initial ATOM tv-out support. This is disabled by default currently,
>(please use the Option "ATOMTvOut" "TRUE" option in xorg.conf to test).
>Adam Jackson (2):
>      Fix connector table mapping for lvtma/kaleidoscope/uniphy.
>      Fix HDMI output setup on DIG1/DIG2.
>Adrian Friedli (1):
>      Add another AGP quirk
>Alex Deucher (80):
>      Revert "IGP: attempt to fix VGA on IGP chips"
>      RADEON: IGP VGA fixes take 2
>      RS4xx: enable the DRI by default on all variants
>      Revert "rotate: fix rotation in conf file."
>      Add quirk for Falcon NW laptop
>      Fix cursor with multi-head and rotation
>      Make sure cursor offsets are properly aligned when using EXA
>      R300: NUM_FPU adjustments for VAP_CNTL
>      ATOM: Add support for UpdateCRTC_DoubleBufferRegisters cmd table
>      ATOM: fix for r4xx cards with no VRAM_Usage tables
>      Add oem quirk for external tmds setup on Dell Inspiron 8600
>      Clean up legacy display watermark setup
>      Clear display priority bits before resetting them
>      Interlaced mode fixups for AVIVO chips
>      First pass at InitDispBandwidth() for AVIVO chips
>      Add quirk for oem x300 card
>      Fix error in driver connector table for powerbook w/ vga
>      Change prim types for exa and textured video to help avoid tearing
>      IGP: Updated quirk handling
>      IGP: attempt to get external TMDS working
>      IGP: fix typo in IGP quirk handling
>      Remove un-needed dac check for single crtc cards
>      RS4xx: Fix up ddc gpio
>      RS4xx DDC fixes take 3
>      Fix VT switching on M6 chips
>      Further cleanup and unification of i2c code
>      Remove reset of 3D scissor registers when using the CP in the ddx
>      PLL adjustments
>      Whitespace cleanup from bucubic merge
>      Additional cleanups and re-arragement following bicubic merge
>      match textured video macro names with other accel code
>      Move CP into a separate struct
>      Move accel state to a separate struct
>      Move DRI to a separate struct
>      Switch to using radeon_drm.h from the drm
>      Convert randr, Xv to a common allocator
>      Bicubic fixes from the last cherry-pick
>      Switch cursors over to generic allocator
>      Switch EXA path back to static cursor allocation
>      fix some fallout from the common allocator
>      radeon: Add quirk for macbook
>      Fix copy/paste error in the last commit
>      more macbook fixes
>      cleanup macbook quirk
>      radeon: Further pll adjustments
>      atombios updates from upstream
>      Workaround to make initial rotation work
>      proper fix for initial rotation
>      man page updates
>      Add support for DDC via atom commands for RV410
>      Refactor atom LCD/DFP output setup
>      Turn coherent mode off by default on DVI
>      set grey levels correctly for temporal dithering
>      Fixups from last DFP/LCD refactor
>      Fix grey level dithering setup
>      Add LCD dithering quirk for macbook pro
>      Check for LVDS on all IGP chips
>      Don't support rotation is accel is not available
>      Get hardcoded edid from the bios for servers that support it
>      Fix up posted logic
>      radeon: fixup shared DDC lines for some rv610 cards.
>      Add quirk for r200 card with the primary dac wired to both ports
>      move 3d state init after RADEON_SWITCH_TO_3D()
>      Only flush IB when direct rendering is enabled
>      Make sure 3D state is emited when switching to 3D IB
>      Fix up US setup after r3xx bicubic merge
>      First pass at tear-free accel
>      Make VSync for EXA and Xv configurable
>      set coherent mode consistently (enabled for now)
>      R200: Remove scissor setup from init3d()
>      Add missing .TP in man page
>      Add randr 1.3 panning support
>      Add another AGP quirk
>      Add support for Apple eMacs
>      Remove mention of MergedFB from server log
>      Remove remnants of unused DDCMode option
>      R3xx-R5xx: switch back to quad rendering for Xv
>      R2xx: switch EXA composite to rects rather quads
>      Pre-avivo: fix FP setup
>      R3xx-R5xx: better fix for xv primitive tearing issues
>Benjamin Close (1):
>      xf86misc extension has been removed, update driver to reflect it
>Benjamin Herrenschmidt (3):
>      atombios: add support for other endians.
>      Fix console switch on R500
>      radeon: Fix PCI usage of 32-bit driver on 64-bit platform
>Brice Goglin (2):
>      use PACKAGE_VERSION_* for radeon version
>      Increase default GART size to 32MB on >=R300
>Bryce Harrington (4):
>      radeon: Fix pasto in connector table setup for vga powerbooks
>      Add AGP quirk table
>      Additional AGP quirks
>      Add another AGP quirk
>Calvin Fong (1):
>      IGP: add support for NTSC tv-out on legacy IGP chips
>Corbin Simpson (21):
>      Add bicubic texture table, as well as the script used to (re)generate
> it. To regenerate, just run "python bicubic_table.py > bicubic_table.h".
> Fixed typos in the bicubic texture tables. Whoops, looks like I'm still
> asleep. Allocate memory for the bicubic filter texture.
>      Upload bicubic filter to card. This was a LOT easier than I had
> feared, to be honest. Upload pixel shader to card for r5xx. This was
> ridiculous. Also it doesn't work yet... Try to get tex coords from the VAP
> to RS properly when bicubic filtering is enabled. Make vertices emit
> properly.
>      More bicubic FP buggies.
>      Finally got the fragment program fully working for bicubic filtering
> on r5xx. Fix constants.
>      Merge upstream changes to vertices, and also add Xv attributes for
> textured video, including bicubic filtering. Oops, made a mistake with
> vertices.
>      Bump bicubic cutoff to R580.
>      Force R580-only for bicubic.
>      Update bicubic tables.
>      Enable bicubic filtering for all r5xx HW.
>      Change floats to uint32_t hex.
>      Switch from 32-bit floats to 16-bit half-floats.
>      Switch to Mesa-style 24-bit float packing.
>      r3xx: Various Xv fixes.
>      Fix indentation on IS_R300_3D Xv code.
>Dave Airlie (42):
>      radeon: drop all use of CPMode.
>      rotate: fix rotation in conf file.
>      rv770 initial support
>      pciid: add radeon HD3850.
>      atombios: use macro to get offset into table
>      atombios: fix typo in mode conversion for panel
>      radeon: make r600 use i2c table lookup for ddc.
>      radeon: add 0x9441 for hd 4870 x2
>      radeon: fix powerpc build
>      radeon: rename radeon_memory to radeon_legacy_memory.
>      radeon: hopefully fix plls for 30" monitors.
>      atombios: set missing pixel clock
>      atombios: oops typo
>      pciid: add HD3300 - 790GX chipset.
>      radeon: always align dstPitch to 64 bytes for textured video
>      r100/r200: EXA misrenders dst == a8 + dst alpha use.
>      radeon_drm.h: remove kernel defines
>      radeon: fix HDMI having DACs in some BIOSes
>      Stability fixes from radeon-gem-cs
>      Don't mix 2D and 3D in the same IB
>      radeon: apply atom quirks to object table
>      radeon: damn you bios, HDMI cannot have DACs.
>      radeon: memset the pll setting function
>      r600: don't idle engine
>      radeon: fixes from Alex for some output engines
>      DCE32: add support for DCE3.2 digital outputs.
>      radeon: add initial rv730 pciids.
>      atombios: fix issue with Clearing of MM_INDEX 0.
>      atombios: correct fix for previous issue.
>      atombios: fixup parsing of TV tables on r500 and r600 cards.
>      atom/tv: fixup the tv out default if bios doesn't specify
>      atom/tv: enable scaler can set the TV output scaler up.
>      atom: spc3 only requires ucEncoderMode for DIG outputs
>      radeon: use latest atombios constants for TV encoders in
> set_crtc_source atom/r600: re-enable TV outputs
>      atombios: re-enable TV dpms
>      radeon: tv timings have crev/frev wrong way around for some reason.
>      radeon: add all new pci ids for rv730/rv710 families
>      radeon: setup 3D engine even when no DRI.
>      atombios/tv: add an option to enable atom tv-out for users.
>      radeon: pass distcheck
>      radeon: update to
>Dennis Kasprzyk (12):
>      Fix texture size, texture filter, vertex offsets, etc.
>      Set helper texture filter correctly.
>      Implement LPR in one instruction.
>      Remove uneeded negations.
>      Heavy optimizations.
>      Remove one constant.
>      Smarter usage of the texture semaphore.
>      Another uneeded SEM_AQUIRE.
>      Move some ALU instructions after the TEX instructions, so we can do
> something usefull while we are waiting for the texture values. Fixed
> bicubic fragment program comments.
>      Fix bicubic fp calculation.
>      Fix typos.
>Egbert Eich (1):
>      Cleanups from rhd port
>Eygene Ryabinkin (1):
>      Catch unsubstituted macro for PCI region base access
>Fabio (3):
>      Fix formatting in man page
>      Remove duplicate ATI in some device names
>      man page updates
>Joseph Adams (1):
>      Add more eMac modes, fix crash in last patch
>Julien Cristau (2):
>      Link with -lpciaccess and -ldrm if needed
>      Bug#7148: Fix manpage formatting
>Maciej Cencora (3):
>      Add needed FP registers, etc. for r3xx bicubic Xv.
>      Add r3xx-specific bicubic filtering code. FP, RS, mostly, everything
> else is shared with r5xx. Enable bicubic Xv on r3xx+. Still some buggies in
> the actual display, but looks pretty good.
>Michel Dänzer (7):
>      Restore versioning of interface for Mesa DRI driver.
>      Only declare local variable when it's used.
>      Call DRM modeset ioctl after the IRQ has been initialized.
>      Make sure video offerlay offsets don't exceed the hardware limit of
> 128 MB. Fix -Werror build.
>      Update GEN_INT_CNTL register value after calling DRM modeset ioctl.
>      Pass base offset into RADEONDisplayVideo() explicitly.
>Owain Gordon Ainsworth (1):
>      radeon: don't call ioctl unless DRI is initialised
>Patrick Haller (1):
>      Fix off by one in EXA composite limit checking
>Pierre Ossman (10):
>      Fix comments for R500 fragment shader to reflect the code.
>      Fix node setup on R300 bicubic pixel shader. It was backwards and used
>      Fix macros so that they are safe to use with expressions as
> parameters. Now that we have a bicubic code path that messes up the pixel
> shader, Make the R300 bicubic shader program a bit easier to understand by
> Change the XV_BICUBIC attribute to a tristate, where the third state Switch
> r200 Xv to use rect lists rather than quads to avoid tearing Improve
> tearing avoidance for Xvideo in two steps
>      Optimise RADEONWaitForVLine
>      Fix scissor setup for Xv
>Ramon van der Stelt (1):
>      Interlaced mode fixes
>Roland Scheidegger (1):
>      clamp tex coords (r100/r200) for textured video
>Wolke Liu (1):
>      Add pci id for FireMV 2400
>git tag: xf86-video-ati-
>1.tar.bz2 MD5: ec6eda1fb77cf2a289e31f2b54edd056 
> xf86-video-ati- SHA1:
> 710827ccbf2ad3660fde9ad9a8aab91b7820af54  xf86-video-ati-
>1.tar.gz MD5: b5c547bfbe7a164845fbda6ab482edba 
> xf86-video-ati- SHA1:
> 71a6d1ff9157c7f8ef44ea5537bf165d76bd6e6e  xf86-video-ati-
Dave, FU8 system, quad core phenom.  ./configure --prefix=/usr, no errors.

Make OTOH, trundles along for a few files and then gets all choked up:
gcc -DHAVE_CONFIG_H -I. -I.. -I./AtomBios/includes -Wall -I/usr/include/xorg -I/usr/include/drm -I/usr/include/X11/dri -DDISABLE_EASF -DENABLE_ALL_SERVICE_FUNCTIONS -DATOM_BIOS -DATOM_BIOS_PARSER -DDRIVER_PARSER -g -O2 -MT 
radeon_accel.lo -MD -MP -MF .deps/radeon_accel.Tpo -c 
radeon_accel.c  -fPIC -DPIC -o .libs/radeon_accel.o
radeon_accel.c: In function 'RADEONEngineInit':
radeon_accel.c:383: error: 'RADEON_PARAM_NUM_GB_PIPES' undeclared (first use 
in this function)
radeon_accel.c:383: error: (Each undeclared identifier is reported only once
radeon_accel.c:383: error: for each function it appears in.)
make[2]: *** [radeon_accel.lo] Error 1
make[2]: Leaving directory `/usr/src/xf86-video-ati-'
make[1]: *** [all-recursive] Error 1
make[1]: Leaving directory `/usr/src/xf86-video-ati-'
make: *** [all] Error 2

What am I missing?


Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
innovate, v.:
	To annoy people.

More information about the xorg mailing list