commit 8cbff7bfa1201faa1e5dbf7019f17c2f4b1d2b7f
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Oct 9 17:42:45 2009 +0200

    Bump to 1.3.0. Update README.

commit bd2145d8fab854fb5617abd09f7517321d2b35be
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Oct 9 17:42:30 2009 +0200

    pm: Comment out currently unused variable.

commit 5c437ecd5d06e1b214d2a309dfed60fadf071829
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Oct 9 17:31:29 2009 +0200

    README.coding: Add rhd_acpi.[ch] and atombios_rev.h

commit 0c8cc053741d7113254776d6c64745ecc5a1256e
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Oct 9 11:25:06 2009 +0200

    i2c: Hardware may need a while to indicate availability to the host
    
    Fixes fdo #24330.

commit 57b97e0fefd7e5a0069573befabdb28ee0094f8d
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Oct 8 17:10:05 2009 +0200

    pm: Improve wording of lowPowerModeEngineClock logging.

commit a1cd56ddd2955ce74cccc4adcd482feda0985b72
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Oct 8 16:59:55 2009 +0200

    pm: Ignore validation when setting negative engine/mem frequencies.

commit 5d5d8e3d8655a7adbc17a548372bc0fad6d67cb7
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Oct 8 16:57:21 2009 +0200

    pm: Use minimum known working frequency instead of default/2.

commit 0a94f8eb56ec0155864858f8aefc4c64d3e1bf77
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Oct 8 16:02:46 2009 +0200

    pm: Do NOT set idle engine clocks to default/2 unless ForceLowPowerMode is used.
    
    Accidentally, we set this by default since commit fedddb7.
    Using different engine clocks has side effects that are not handled yet.

commit eebdbf0a9723c79975015354830550f7dfb57509
Author: Egbert Eich <eich@freedesktop.org>
Date:   Thu Oct 8 09:15:32 2009 +0200

    Improve DAC load detection on RS780.
    
    Increase delay before reading out DAC autodetect status.

commit 422ac06b69cfcbfbaa802fdc916d3b87f40eeb41
Author: Christian König <deathsimple@vodafone.de>
Date:   Thu Oct 8 00:04:01 2009 +0200

    Silence audio stream option.
    
    This patch adds an option to play silence instead of playing no sound at all.
    This fixes some issues where a receiver needs 1-2 seconds to adjust to a new audio stream.
    This option can be enabled by setting AudioStreamSilence in xorg.conf or by xrandr.

commit 3cef2a6553359b063b760a2f436ce84365c60682
Author: Egbert Eich <eich@linux-21vw.site>
Date:   Wed Jul 1 14:06:27 2009 +0200

    Improve test for disabled differential clock driver.

commit 392a13923eb4d44a8bdb204922230e126b188fae
Author: Egbert Eich <eich@freedesktop.org>
Date:   Wed Oct 7 12:23:56 2009 +0200

    Properites: Get HW for Backlight setting on every query.
    
     * Backlight brightness may be changed from underneath (ie. thru hot keys).
       Thus the HW needs to be reprobed every time.
     * This patch adds ACPI backlight support also for AtomBIOS based mode
       setting.

commit f695445b386050de346211baec5e487813187cfd
Author: David Morrison <dave@bnl.gov>
Date:   Tue Oct 6 11:13:26 2009 -0700

    LUT: Fix syntax error in 59085c4a

commit 59085c4aa0f420a56699c0a77aa3fd6883b41efe
Author: Yang Zhao <yang@yangman.ca>
Date:   Tue Oct 6 10:47:38 2009 -0700

    LUT: Fix RHDLUTCopyForRR() not actually modifying LUT

commit 8b3561d2aabeed34c77d9845e2f9814de8375bed
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Oct 6 19:30:32 2009 +0200

    ID: Added quirk entry for HIS Excalibur Radeon XT1650 Pro IceQ 256M
    
    This card reports 2 DVI ports but has only 1 VGA and 1 DVI.
    Also HPD isn't working on the DVI port.

commit 0a79a28c5552a437e4bc4876a2aac9ce8c062485
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Oct 6 19:29:44 2009 +0200

    PM: Removed unnneeded define.

commit 4758ba31d49dbafb1e8c9e19aad84d41183c7f19
Author: Egbert Eich <eich@freedesktop.org>
Date:   Mon Oct 5 22:19:27 2009 +0200

    Add ACPI controlled Backlight support for Linux.
    
    ACPI Backlight support is system specific and only implemented for
    Linux so far.

commit be7216fca954396d92b94335ccb18d22c354c195
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 30 15:46:46 2009 +0200

    randr: Select virtual large enough for typical dual-monitor situations
    
    Unless we're able to shrink/enlarge FB on the fly (TTM etc.), allocate
    large enough (TM) virtual size for general use cases.

commit b9648ff4f60a55e9997afc574c91b1f92645f7c2
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 30 15:41:58 2009 +0200

    RV740: Choose shadowfb by default.

commit eb1fc6c0070077a23a4eb20c934e2743b6661b9b
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 30 12:48:55 2009 +0200

    Fix missing parenthesis.
    
    As ! has a higher precedence than &, this commit changes the semantics. This
    could fix some obscure bugs, but could also regress.

commit 136dbca77332bf119768b0719ed069d4fa1c74a9
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 30 12:16:26 2009 +0200

    Disable DRI by default if shadowfb is explicitly selected.

commit 209aa4c76330d15c387185f61be008b412aae4b3
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 30 12:04:05 2009 +0200

    RV740: disable DFS and Composite
    
    Small transfers seem problematic on all rv740 chips, and composite cannot be
    disabled per size, so until we can find a solution, disable them.

commit 21f033c1f812598d5e80d5bc9462807244f0d037
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Sep 7 16:06:04 2009 +0200

    Add quirk table entry for Wyse R-Class.

commit ee508b375ba6a1fe132cb3bd8cf461445c814ea8
Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
Date:   Sat Sep 19 15:15:35 2009 +0200

    Hide README, radeon.man non-updates if --enable-shave
    
    If configured with --enable-shave (default), hide the updates
    to README, radeon.man if they result in zero changes.
    
    If they result in changes, the changes will still be shown.
    
    On --disable-shave, the commands generating README.new and radeon.man.new
    will always be shown.

commit b7a50a6f00bb6d207f5240f00cf83fbbd56bbf60
Author: Hans Ulrich Niedermann <hun@n-dimensional.de>
Date:   Fri Sep 18 23:23:32 2009 +0200

    Fix RHDRegWrite macro invocation breakage
    
    This appears to be an obvious fix to a bug introduced with bab7c5f9.
    
    Neither ignoring the blue component nor discarding the calculation
    does made sense to me, and the obvious explanation for that would
    be a misplaced parenthesis.

commit ae41ab2fe49ca456c4fb23859af8683c32c2429c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 16 17:22:28 2009 +0200

    Add quirk table entry for ATI Radeon X1550.

commit 37ccdde4fef03f132ed2cd5cb6797f83f2b7f41e
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Sep 14 14:21:35 2009 +0200

    Add ugly hack for uint32_t to rhd_pm as well.

commit d9d0cecbfa54a58c2553e1750ee8124e178679b0
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Sep 14 13:16:46 2009 +0200

    Fix typo in non-DRI case.

commit ca1e34f85387bdf3eb1727d2fb628e4774f893a1
Author: Luc Verhaegen <libv@skynet.be>
Date:   Tue Sep 8 17:22:57 2009 +0200

    Revert "Fix softlocks on rs690. Idle commands have to be flushed to be of any use."
    
    This reverts commit d8329927aaac5f2d4949785951326ebc782bc420.

commit d934165ce11d5c8abfe577a72f57cc97c4a6e734
Author: Luc Verhaegen <libv@skynet.be>
Date:   Tue Sep 8 17:20:41 2009 +0200

    Revert "More idle/flush swaps."
    
    This reverts commit b3f16ffdeb95fcd07f773756a563a5873647a760.
    
    When looking at this code in depth, it is clear that these changes make the
    sequence of commands useless.

commit b3f16ffdeb95fcd07f773756a563a5873647a760
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Sep 8 16:53:50 2009 +0200

    More idle/flush swaps.

commit d8329927aaac5f2d4949785951326ebc782bc420
Author: Marc Dietrich <marvin24@gmx.de>
Date:   Tue Sep 8 15:32:38 2009 +0200

    Fix softlocks on rs690. Idle commands have to be flushed to be of any use.

commit b00c0a75a97136aaae887580a51acefc25ce7c4b
Author: Yang Zhao <yang@yangman.ca>
Date:   Sun Sep 6 21:42:40 2009 -0700

    LUT: Make sure LUT_RW_INDEX is reset to 0 before loading LUT

commit 57916f27c08122b6aaea0da44c0ba21634706dce
Author: Yang Zhao <yang@yangman.ca>
Date:   Sun Sep 6 15:53:57 2009 -0700

    LUT: Use DC_LUT_30_COLOR for saving/restoring LUT

commit bab7c5f97ddcd23772e6fdd66084f791c30218cd
Author: Yang Zhao <yang@yangman.ca>
Date:   Sun Sep 6 15:31:15 2009 -0700

    LUT: Change internal functions to reflect hardware
    
    Existing implementation of LUT setter uses a LoadPalette()-like
    function signature, where as the actual hardware is more accurately
    reflected by RandR style gamma setter.  On a RandR gamma adjustment
    call, this results the following sequence of frivolous conversions:
    
      (u16*,u16*,u16*) -> LOCO* -> (u16*,u16*,u16*)
    
    New implementation changes the LUT setter to accept a triplet of CARD16*,
    and move the LoadPalette()-specific logic to RHDLoadPalette().
    
    A new rhdLUTSetRow() function is added, which deals with setting only
    specific rows of a LUT.  LUTxSet() previously made bad assumptions in
    these cases.

commit 27cfbaa3c34844771b5a4262a5deaf71b79cb6bc
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Sep 2 16:45:12 2009 +0200

    atombios.h prohibits -std=c99 due to unnamed unions

commit fa6312784e3a03afb17ec1ea5e0e1912ca34d2b9
Author: Marc Dietrich <marvin24@gmx.de>
Date:   Fri Aug 28 09:07:34 2009 +0200

    silence some compiler warnings
    
    This patch silences some compile warnings
    mostly comming from C++ style comments.
    
    I left the Atombios stuff untouched.

commit f7ad9380b7e5af6b0c85a5423ab5b7a31f08fd8c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Sep 1 16:17:11 2009 +0200

    Fix segfault due to missing Card NULL check.

commit f047271abb94226df0076f3715c960e5b3c45b06
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Sep 1 14:54:04 2009 +0200

    Also combine outputs if only the second of them has .dual set.
    
    Additionally, sometimes the VGA output is listed before the DVI-I.
    Previously these outputs were not combined.

commit 9ca0fc45ceee43003a983176301edde0bcaad2fe
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Aug 28 17:12:06 2009 +0200

    Add atombios_rev.h to files list.

commit fe7dbc2484edcf92944c11dde50b50e5b21f0bad
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Aug 18 12:28:54 2009 +0200

    Consider Backlight property to be a known good property now.

commit 78b6ed065aa6bc9bd66a20665daad9849e3696e3
Author: Yang Zhao <yang@yangman.ca>
Date:   Tue Aug 25 14:52:21 2009 -0700

    Put back logging of PCI IDs not in database during probing.
    
    Revert of 45e733556671e78adbb2afafd6b98a461bb29825 with modifications.
    That was the only place were users are informed of the maillinglist, so
    removing it was a bad idea.

commit 45e733556671e78adbb2afafd6b98a461bb29825
Author: Yang Zhao <yang@yangman.ca>
Date:   Mon Aug 17 16:39:34 2009 -0700

    Do not log about "Unknown card detected" when AtomBIOS is available
    
    There's a lot of confusion about this message; time to kill it now that
    AtomBIOS modesetting is significantly more stable.

commit 10133f5d04c2e709c7cf2ed3d6dcd16b13b96607
Author: Yang Zhao <yang@yangman.ca>
Date:   Mon Aug 17 16:27:42 2009 -0700

    Use DRM and EXA by default on r6xx/r7xx
    
    Manpage updated accordingly.

commit d44b359c7b12236f0045734bcc42de397409ab9f
Author: Yang Zhao <yang@yangman.ca>
Date:   Mon Aug 17 16:21:52 2009 -0700

    Fallback to ShadowFB on r6xx/r7xx if DRI pre-init fails

commit ed98d7311d021849993ac764b2caffd6ddceeb2a
Author: Yang Zhao <yang@yangman.ca>
Date:   Wed May 6 14:22:13 2009 -0700

    Make EXA the default AccelMethod on r5xx
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 865215c800f50d8718544acf2af222c5f2ef777a
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Aug 14 12:44:50 2009 +0200

    Non-existent AtomBIOS command tables should issue a warning only.
    
    This is not fatal and sometimes even anticipated.

commit be2c65fa35e8f9ae04e799f43400822478d764d2
Author: Alex Deucher <alexdeucher@gmail.com>
Date:   Wed Aug 5 01:01:07 2009 -0400

    rs880: enable accel

commit 62e60e31db474e69dbb2dcee2838cb58039c1be6
Author: Alex Deucher <alexdeucher@gmail.com>
Date:   Mon Aug 3 16:02:43 2009 -0400

    r7xx: add some new pci ids

commit d0ff1c5823b3c186dd80fb65105b62bff28630e6
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 30 17:16:14 2009 +0200

    INSTALL: xorgversion.m4 is obsolete for a long time already.

commit 2d73df84b71378bc427054397d936e66ceea6769
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 27 19:08:11 2009 +0200

    Fix logic bug in sideport memory detection.

commit f6c57ac9630d1bd8d4e417ae1a05a1abd1573520
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 27 18:11:46 2009 +0200

    MC: Major overhaul for RS600, RS690, RS760, RS780, RS880.
    
    Use same registers and register spaces as radeon now.

commit 83ed13f303538197ab73a02cdb68b4624d15d8e2
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 27 17:32:59 2009 +0200

    Rename registers in PCIE space.

commit b5911fd908efe63933f620ec39511444c5f4964f
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 27 15:56:40 2009 +0200

    RS690 MC idle: Wait for system idle instead of sequencer idle.
    
    Inspired by radeon code.

commit 050f523309063d8e836169014dd4f6a991f555b7
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 27 11:48:31 2009 +0200

    Typo.

commit da5b9c2fb0338e616f5376b6e40b4531fa3d7e1c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Aug 3 17:17:44 2009 +0200

    Fix autoconf issues.
    
    Apparently, autoconf 2.64 has "interesting" issues with spaces in double
    quoted strings, as used in configure.ac. Change that to parantheses.
    Based on a patch by Jens Lody <radeonhd@jenslody.de>.
    
    Additionally, autogen.sh didn't --force so far.

commit 9cabfb076b874f9ab17bb4d219ac492d55e77abb
Author: Yang Zhao <yang@yangman.ca>
Date:   Mon Jul 27 23:08:42 2009 -0700

    Xv: Make color space selectable on r6xx/r7xx
    
    Add functionality to switch between Rec.601 and Rec.709 color space
    conversion constants.
    
    Default behaviour is to use Rec.709 for videos 928px or wider,
    and Rec.601 otherwise.  Force a particular color space by setting the
    new XV_COLORSPACE attribute to 1 (Rec.601) or 2 (Rec.709).  A value of 0
    means do auto-detection.
    
    man page updated accordingly.
    
    This is a follow-up to bug #22901.

commit ca21df5b3583c64cd9f78d449b689d4f48158002
Author: Dave Airlie <airlied@redhat.com>
Date:   Tue Jul 28 15:22:41 2009 +1000

    radeonhd: change to using ABI version check

commit d389503eace03c0d8adf2028ade099b7a7f38a8a
Author: Dave Airlie <airlied@redhat.com>
Date:   Tue Jul 28 13:39:32 2009 +1000

    radeonhd: further RAC/resources changes

commit e688cffa33ececfef8630c05380f93d9e38b3edf
Author: Dave Airlie <airlied@redhat.com>
Date:   Tue Jul 28 13:32:35 2009 +1000

    radeonhd: update for resources/RAC API removal

commit d19a10b753d3a235cda21246c1b1721673fe1061
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jul 24 16:01:30 2009 +0200

    Work around ParseTable returning CD_SUCCESS if the command table doesn't exist
    
    Typical example: GET_MEMORY_CLOCK doesn't exist, and some random value that
    happens to be stored in the call struct is considered to be the correct value.

commit 3d031afba8c23b7099d32b97a16ccdf06a906fdd
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jul 24 15:30:39 2009 +0200

    Validate current clocks - some AtomBIOSes provide broken values.

commit eabbde4e4cc8fa2bd9f3f1f347e557b6b1210f0d
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 18:52:39 2009 +0200

    Fix C++ comments warnings in reverse engineered AtomBIOS.

commit baa8d6e2de33b0b4f7276af86b701bf2f16452bc
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 18:48:42 2009 +0200

    Add check for -Wno-unknown-pragmas in configure.ac.

commit 6f32189092c30d05452063367a440a59ae920f50
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 18:34:19 2009 +0200

    Import known good chip settings from AtomBIOS.
    
    Include reverse engineered AtomBIOS addon header.
    Pragmas in the header are needed for AtomDis.

commit 0684714cc2542b37248f746c20f561b5c577b2c4
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 16:54:25 2009 +0200

    Remove warnings in rhd_pm.c on 32bit.

commit e5e54f76d74bb86ce9436448cee27354e58e9b5d
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 16:51:27 2009 +0200

    Consolidate structures in rhd_pm.c and rhd_atombios.c.
    
    Also nuke copy routine for power states.

commit 9c66f5fbd43a42db4f4843abafd0e103551171a8
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 15:44:22 2009 +0200

    Improve chip limit validations. Again. 2.

commit eed4bed6ab3e79eb1510b86332a073496fe61761
Author: Javeed Shaikh <syscrash2k@gmail.com>
Date:   Thu Jul 23 13:10:56 2009 +0200

    Xv: Fix YUV scaling and use Rec709 coefficients
    
    Described in fdo bug 22901. Coefficients are now based on the color space used
    for HDTV, which seems more appropriate.
    Also fixes some off-by-one bugs.

commit 7c2df35a8fa7bc836f1444a508ac3ab5dcab15e9
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 12:46:40 2009 +0200

    Improve chip limit validations. Again.
    
    Probably will take a few more iterations until we get it completely right.

commit 8f0f6217feb198200c60a0e11fb332a9cae28dd5
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 12:40:37 2009 +0200

    Improve output for powermanagement.

commit 8f3f3647aad1613c41ec4b5bd030e571b74b7cf2
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 12:31:09 2009 +0200

    Improve chip limit validations.

commit e9cd2d257f304041fd1a127a58aeea0415d984bb
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jul 23 11:51:00 2009 +0200

    Typo.

commit 98f46c6abebdc63b8be3d2c814da1ada9aa8ddef
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Jul 22 18:41:08 2009 +0200

    Add code fetch chip clock and voltage limits and validate them.

commit d9288b43ecd639713e64aa435943697c5e1c9c8c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Jul 21 17:43:32 2009 +0200

    Consolidate prefixes for ATOM_*.
    
    Some had ATOMBIOS_, some had no prefix.

commit fedddb78ab01e3972854ba6b115ea653fa8a4b7c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Jul 21 17:08:37 2009 +0200

    Rework powermanagement: Introduce the notion of levels, change interface.
    
    Old style didn't scale with new requirements.

commit b1be3143266f8b6a89dfe635321f16797601f1bb
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Jul 21 12:10:53 2009 +0200

    Rewrite to use structured data for power settings.

commit 5995c731da030631e56020e8cc4dcdcb92f5f935
Author: Egbert Eich <eich@freedesktop.org>
Date:   Wed Jul 22 13:15:49 2009 +0200

    Clean up most warnings.

commit 2e2e4d118e497cfc50f0bc23741236a89d9798f3
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Jul 21 18:04:07 2009 +0200

    Add better fallback heuristics for acceleration methods.
    
    - R6xx and up don't support XAA: fall back to EXA.
    - If DRI is not compiled in on R6xx and up fall back to Shadow_FB.

commit de3f80bc6ec2adaab40912fee92fe5f38564d1fe
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Jun 30 18:54:18 2009 +0200

    Reenabling AtomBIOS controlled backlight.
    
    This code used to be commented out because it triggered some crashes.
    It doesn't crash for me (tested LVTMA nor DIG). Let's wait and see
    what reports will come in and check what the issue is.

commit 1ff8649877a8d75b2614b40f90da60e3690fc88d
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Jun 30 18:47:54 2009 +0200

    Move test for backlight support enabled to proper place.

commit e4d5f972358ffbd115eafce00fdc578ed2a98cc6
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Jun 30 18:46:34 2009 +0200

    Spelling fixes.

commit a006e56b0b845028780a25d905cd15eb94b34fd6
Author: Egbert Eich <eich@freedesktop.org>
Date:   Tue Jun 30 18:45:26 2009 +0200

    Remove superfluous function arguments.

commit b22e7a21a589b5205ad61a8cf39a61c8e5ca5791
Author: Peter Hutterer <peter.hutterer@who-t.net>
Date:   Thu Jul 16 13:27:21 2009 +1000

    radeonhd: Update to xextproto 7.1 support.
    
    DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers
    need to include dpmsconst.h if xextproto 7.1 is available.
    
    Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>

commit bc42c63d0cf7756a9d31c16d68d1c33a9e225b83
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jul 6 19:00:06 2009 +0200

    Add shave (nuking excessive libtool output).

commit 09a62f9426ccf5c6cbcc19cb9fd7ba66c3231002
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Tue Jun 30 12:57:03 2009 +0200

    Fix build with --disable-atombios

commit 3e0d3aab927db2def6375b11fa97e788d1b6dcec
Author: Luc Verhaegen <libv@skynet.be>
Date:   Mon Jun 29 21:31:43 2009 +0200

    DRI: Fix VBlankInterrupt setting when not using RandR1.2.

commit c7168b6e835fdf7f2837913ed8580f1911135fdd
Author: Luc Verhaegen <libv@skynet.be>
Date:   Mon Jun 29 21:03:05 2009 +0200

    Cursor: fix logic error in cursor visibility checking.
    
    Cursor->(X,Y) were always null and never set until Crtc->(X,Y) were within
    CURSOR_MAX_(WIDTH,HEIGHT) boundaries. This only affected xf86Cursor code.

commit 645a0a0a6e5e57a4bb77c3b9c7d20b58d4c9ec28
Author: Roland Scheidegger <sroland@tungstengraphics.com>
Date:   Sun Jun 28 17:00:20 2009 -0400

    R6xx/R7xx: fix pixel centers
    
    Make sure we are using GL mode (centers at 0.5) rather
    than D3D mode (centers at 0.0). This also fixes fdo
    bug 21963.

commit afc94a3b398340768e9e6aac07a1609e8302dd84
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 26 15:34:43 2009 +0200

    Add Developer's documentation. Or better: Some start of.
    
    Initiated by Rafal, who had to ask Dave, John, and Jerome for quite some
    stuff, and others might benefit from that.

commit 0624f7e6a9c5fff44c8e50e0df6a97835ad3484f
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 26 12:18:39 2009 +0200

    Add quirks table entry for ASUS EAH3450.

commit fbc7fbfa1e7244622adfbd8850a742818e18df72
Author: Christian Koenig <deathsimple@vodafone.de>
Date:   Wed Jun 10 00:43:11 2009 +0200

    R700 HDMI audio support.
    
    Setting just one more bit gets audio going on R700.
    This time the tanks for debugging this goes to Jaren Peterson.

commit 47a052a44f7566431308c5441539b65bb3dc6600
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Thu Jun 18 09:57:31 2009 +0200

    PM: Fix printing units and variable typo in memory code

commit 8a3fa92b6fdcd86b502973c947e5aed1bd3e1ada
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed May 13 16:36:54 2009 +0200

    randr: Reformat AtomBIOS parameter correctly when setting.

commit ebcac22cadb28132a4133426ff9c5d662f9ff3e6
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed May 13 16:08:52 2009 +0200

    randr: XA_STRING property values are not necessarily 0-terminated.

commit 87dc50fbfa9a284f6387356b4840de6598de8139
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jun 15 16:35:19 2009 +0200

    randr: Export EDID property correctly.
    
    Wasn't exported because rhdRROutputSetProperty() is called by RandR layer, and
    the function didn't know about this property.

commit f7260fc38a50b50395ff21ff76d568d7761677b1
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jun 15 16:32:43 2009 +0200

    randr: Add AtomBIOS property for on-the-fly setting. Experimental.
    
    Currently disabled by default.
    To enable set ENABLE_PROPERTY_ATOMBIOS to 1 in rhd_randr.c.

commit aa1a56fe12b76abe8d483fdb0b6d0858c43a5d90
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed May 13 15:44:51 2009 +0200

    Refactor AtomBIOS option handling.

commit 4b4a439ef35af340b9c8e48e3760e375b07f873c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Jun 10 12:23:55 2009 +0200

    Move setting of Private->Modes to correct place.
    
    Was done for LVTMA only before.
    Based on a patch by Rafał Miłecki. Fixes xorg bug #21877.

commit fc31eb65edf6ffac969eb33a7c588dad7c93b7f4
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Tue Jun 9 18:26:30 2009 +0200

    PLL: RV620 PLL shutdowning fix (sometimes we need unused PLL running)
    
    This makes RV620PLLnPower follow AtomBIOS method, fixes bug #18016

commit c83bd5c43832ed5602ee75db74bbe9ca06922d16
Author: Matthias Hopf <mhopf@suse.de>
Date:   Tue Jun 9 15:34:14 2009 +0200

    randr: Remove useless error messages when fetching unknown/invalid properties.

commit 9e8608c4ee55ca7798c3c120938d2ecf1f7e9f00
Author: Matthias Hopf <mhopf@suse.de>
Date:   Mon Jun 8 13:45:21 2009 +0200

    Adapt for nuked xf86LoaderReqSymLists() in Xserver 1.6.99.1.

commit 1103cff7f79e3212af2dcf3e0bf684006c9b79e4
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 5 19:32:55 2009 +0200

    Add quirk table entry for Asus F3JR.

commit 5dd9d814c8a9a35fbf5556d88297eb5e934731d7
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 5 19:22:27 2009 +0200

    Initialize fresh connector object correctly in AtomBIOS case.
    
    Otherwise connectors without DDC object would get DDC0.

commit cca9c0b2766139e225208c6d1e39c56f865480b2
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 5 18:36:16 2009 +0200

    Add AtomBIOS based static powermanagement enablement
    
    Also add dynamic clock gating on pre-R6xx chips, but only if AtomBIOS is
    available.

commit 086e6c4aae8b9d1f077e39b35c1e15b92df0903c
Author: Matthias Hopf <mhopf@suse.de>
Date:   Fri Jun 5 16:49:45 2009 +0200

    Add some more PCI Ids, improve RV740 handling, fix descriptions.

commit 1eec42729e048d7bb9be7fbf50f49a0573093c4d
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Wed Jun 3 19:56:27 2009 +0200

    AtomBIOS: add pixel clock's ucFracFbDiv to debugging messages

commit b8533a98448f8dc2d67137e02a180430def17e0b
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Wed Jun 3 19:27:24 2009 +0200

    RandR: add returning HDMI property value

commit efa0b5475517e4252c5a97389ee5612ed7da453f
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Wed Jun 3 19:03:48 2009 +0200

    RandR: fix return in DigPropertyControl
    
    Currently for all properties (coherent, backlight and hdmi) we
    always return true. This leads to exporting not supported proprs
    to randr.

commit c559887b9663531a8aac526031e6eb4ca7c48efd
Author: Matthias Hopf <mhopf@suse.de>
Date:   Thu Jun 4 15:13:27 2009 +0200

    Add quirk entry for 0x7146:0x174B:0x0940. Bug #21947.

commit c6f607fd6cc40d6f5f9ebbecd1f1e5152d5d0a9c
Author: Yang Zhao <yang@yangman.ca>
Date:   Wed Jun 3 12:33:59 2009 -0700

    Fix incorrectly reporting XAA/EXA init as failing for r5xx
    
    Oops, missed some braces.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit d89f5d870a816291620d18439eebae79557bde53
Author: Yang Zhao <yang@yangman.ca>
Date:   Thu May 28 21:27:31 2009 -0700

    Raise warnings when 2D acceleration initialization fails
    
    Raise a warning when 2D acceleration is diabled during initialization
    for whatever reason.
    
    Also update the R600 CS unavailability warning to point out requirement
    of DRI instead of cliaming lack if implementation.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 1d611211d8814c462e658c7cb8771ea98c9a7bb6
Author: Yang Zhao <yang@yangman.ca>
Date:   Sat May 23 11:59:38 2009 -0700

    MC: Remove duplicated definition of SRBM_STATUS
    
    Use SRBM_STATUS from r600_reg_r6xx.h and remove old definition in
    rhd_regs.h.
    
    Show names of bits are actually being checked in R600MCWaitIdle() instead
    of using magic number.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 792c736887bf99f8631e385fa35cd0ec1e1b1713
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed Jun 3 19:14:05 2009 +0200

    Fix switch construction in LVDSPropertyControl

commit 23653f8d3abe08da04feb7024d643071f09592aa
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Fri May 22 11:39:01 2009 +0200

    PM: add memory downclocking. Commented out for now.
    
    Currently we don't call RHDPmSetClock directly after rhdAllIdle(...) which
    may be needed. It works on my M82 but we need more testing.

commit ebabbe7220cae9aa281b097ad826badeb37a5e83
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Fri May 22 11:24:12 2009 +0200

    PM: get rid of EnableForced field
    
    We can use ForcedEngineClock to determine if we should change engine clock.
    It's set to 0 on init and some bigger value if we decide to downclock. It's
    unnecessary and would be pain in future to keep fields like ForceEngine,
    ForceMemory, ForceLanes, etc.

commit d84490c8ce93f51815cd81c3a8624eecf9cf396a
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Fri May 22 11:10:24 2009 +0200

    PM: add RHDGetDefaultMemoryClock

commit b452d15fcf14401f88ef36bfa8eaa1fede1f15ab
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Fri May 22 10:48:12 2009 +0200

    PM: fix broken engine clock setting logic
    
    Block of code detecting RHDGetDefaultEngineClock failure was placed in wrong
    place. Additionaly Pm->Stored wasn't set to FALSE on init.

commit 08acc05a4c252ff5c55ed1c8f55f106aa6c68546
Author: Matthias Hopf <mhopf@suse.de>
Date:   Wed May 20 18:14:29 2009 +0200

    modes: Fixup broken panel modes if possible
    
    Panels only, and if we can do so only. Otherwise we might have no chance of
    driving the panel at all. Inspired by commit 57a02b50c from the intel driver.
    Fixes Novell bug #472013.

commit d5844f5ba6708b186b2914ddc7cb39c3833570b8
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Tue May 19 11:36:41 2009 +0200

    PM: Save and restore engine clock on VT switching

commit 2644e121f404c9112293debe19dc5241d3b25326
Author: Rafał Miłecki <zajec5@gmail.com>
Date:   Tue May 19 09:38:15 2009 +0200

    clean power management code
    
    This patch moves PM initialization to separated function, creates
    struct rhdPm and stores ForcedEngineClock.
    
    It will be followed by saving & restoring engine clock on VT switching.
    
    --
    Rafał Miłecki

commit 5a0b48dae2284f8d2bf6f8c9584f60177ee74ce6
Author: Christian Koenig <deathsimple@vodafone.de>
Date:   Wed May 13 22:20:09 2009 +0200

    Initial RS690 HDMI Audio support.
    
    Based on a patch from Christiaan van Dijk, but stripped down the bare minimum.

commit 23d25fe7ed0ad6d519672439484cb39653785848
Author: Yang Zhao <yang@yangman.ca>
Date:   Thu May 14 00:42:04 2009 -0700

    Cursor: optimize lockCursor()
    
    It is not necessary to use RHDRegMask() when manipulating D*CUR_UPDATE.
    Using RHDRegWrite() instead saves 2 register reads per mouse operation.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 08461642370b3315cf3412c3a35bf5f1f2db251a
Author: Yang Zhao <yang@yangman.ca>
Date:   Thu May 14 00:30:07 2009 -0700

    Cursor: clean-up: eliminate Width and Height from rhdCursor struct
    
    Hardware always expects a 64x64 cursor image, so there is no benefit to
    explicitly storing cursor size.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit dd287015dcf8ee74de35bd14c8a1a7d63cabdbc7
Author: Yang Zhao <yang@yangman.ca>
Date:   Fri May 8 16:09:56 2009 -0700

    Cursor: refactor non-RandR cursor code
    
    The non-RandR cursor calls now call into the rhdCrtc* cursor functions.
    
    displayCursor() has been eliminated entirely.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 4be5f7152f71c292f16b6e30c59c07b282ea4772
Author: Yang Zhao <yang@yangman.ca>
Date:   Mon May 4 13:38:27 2009 -0700

    Cursor: Fix remaning corruption cases
    
    Port of f668cc06cd7f338888a7dce1507026af0e9e36ad to new code
    
    Found by Alex:
    Apparently the cursor image cannot end on a multiple of 128 pixels.
    Also, for panning, the end of the cursor image cannot extend past
    the end of the viewport.
    
    Signed-off-by: Yang Zhao <yang@yangman.ca>

commit 669e0befdf73a6e1502b0f499a4a9d2122929854
Author: Yang Zhao <yang@yangman.ca>
[--snip--]
