commit 7be2dce975d60d68dd962d4928b2971cd9e011afAuthor: Dave Airlie <airlied@linux.ie>Date:   Wed Oct 4 01:01:48 2006 +1000    update configure.ac for new 6.6.3 releasecommit 6831297d79623d55c57cd7cd8ff90f08b8c68432Author: Dave Airlie <airlied@starflyer-ubuntu.(none)>Date:   Tue Sep 12 03:33:35 2006 +1000    radeon: fallback render repeat for XAA, this fixes corruptions seen on ubuntu        This isn't the correct fix, we should just fix the repeat case, but for a    stable release this is much better idea    (cherry picked from 84e574986f61c4f250d6ecb938bab20847618906 commit)commit ddaacf78455290ffc31b41cd809b03ffbeadd69bAuthor: Ronald Wahl <rwahl@gmx.de>Date:   Wed Oct 4 00:55:43 2006 +1000    Hook up DFP power to DPMS events.    (cherry picked from f96ddd32d3a9e64dc0adec5c50386d6db5b182ee commit)commit 14f6d58d4b5a277a68d9d8ddee7f79be2179f34fAuthor: Roland Scheidegger <rscheidegger_lists@hispeed.ch>Date:   Sat Sep 30 15:14:42 2006 +0200    radeon: Check ROM more thoroughly before treating it as an x86 BIOS.        This prevents the ROMs on some Mac cards getting treated as x86 BIOSen,    resulting in garbage values.    (cherry picked from c979c7f87fc449cc9631820101edff82c3013e66 commit)commit 4921228fb76ec26249aa0f6f0f1ddebf26661b82Author: David Woodhouse <dwmw2@infradead.org>Date:   Sun Oct 1 18:22:14 2006 +0200    Fix radeon vs fbdev pitch calculation.        Fetch the correct displayWidth from fbdevhw after setting the mode.    (cherry picked from 145c716de6d42778e447581578f3e4db13c9a653 commit)commit 9fa176c7989030e7340cc9b85d0b6a7b34303091Author: Alex Deucher <alex@samba.(none)>Date:   Tue Sep 19 13:58:17 2006 -0400    Provide all resolutions (instead only the largest one) in MergedFB clone    mode without config file (Henry Zhao)commit e742aeb28c7d9d6e75932c408bcc7c44af52e303Author: Alex Deucher <alex@samba.(none)>Date:   Tue Sep 19 13:45:33 2006 -0400    fix large negative refresh rates with mergedfb (Henry Zhao)    fixes bug 6966commit 9c0c805704f1df1b66adc2a6b9313597ae5b86ccAuthor: Dave Airlie <airlied@linux.ie>Date:   Mon Sep 18 19:41:21 2006 +1000    radeon: fix rn50 memmap        This should fixup the rn50/m6/m7 memory map problem that has been reported    since BenH's memory mapping changescommit 25fa71ca0a24190cc8be2e80769443e0e09303b3Author: Henry Zhao <henryz@ati.com>Date:   Mon Sep 18 03:15:54 2006 +1000    Power up DAC around check for CRT connectednesscommit 39fcac431eba60adf2d4e357cfc811dbae36a4e1Author: Henry Zhao <henryz@ati.com>Date:   Mon Sep 18 03:09:43 2006 +1000    Bump up clock rate for ES1000 to allow higher resolutionscommit 7d9e2b450f0fa42553156f57572a6dab70a0ea13Author: Dave Airlie <airlied@starflyer-ubuntu.(none)>Date:   Tue Sep 12 07:04:23 2006 +1000    radeon: this stops mergedfb working on my ATI M7 development board        I've no idea what this is for, it came from an ATI codedrop and it may    fix some laptop or production system, but it looks to me more like a dirty    hack than a true fix, as I have an M7 which I need to specify CRT for the    second head and it breaks mergedfb for me.commit ad1778eb1eb59de14ec309b99131931cbccf64e9Author: Michel D盲nzer <michel@tungstengraphics.com>Date:   Sun Sep 10 23:32:44 2006 +0200    Revert "* [ Bug 6966 ] large negative refresh rates reported"        This reverts caaed927a07ffbac68b08246185ef93c1e7bb98c commit.        See https://bugs.freedesktop.org/show_bug.cgi?id=8137 .commit 6e92981b024b59c2237f949321f6b2591dada519Author: Tilman Sauerbeck <tilman@code-monkey.de>Date:   Sun Sep 3 21:11:10 2006 +0200    Removed the warning re. the r300 DRI driver status.commit caaed927a07ffbac68b08246185ef93c1e7bb98cAuthor: Henry Zhao <henryz@localhost.localdomain>Date:   Thu Aug 31 13:27:45 2006 -0700    * [ Bug 6966 ] large negative refresh rates reported    * [ Bug 6309 ] radeon/r128 fails to read hsync/vsync rates when rang    *              descriptor is missing    * [ Bug 5473 ] Blank screen with Radeon Mobility X700 (Acer Ferrari 4005)    * Reboot should not be needed for external CRT to function when connected      later on Ferrari 4000    * Modes need to be sorted, and sorted correctly    * Provide all resolutions (instead only the largest one) in MergedFB clone      mode without config file    * Bump up clock rate for ES1000 to allow higher resolutioncommit 6d2028b84ff2b82c849b74776258f8e52c952e36Author: Dave Airlie <airlied@linux.ie>Date:   Fri Aug 25 10:05:44 2006 +1000    and of course it takes two commits to do a simple commitcommit e90c6feddbcd719fda3bbc3d48fd55045e9d5adcAuthor: Dave Airlie <airlied@linux.ie>Date:   Fri Aug 25 09:59:58 2006 +1000    add changelog hook from nv drivercommit 1e06212f792d1ed1f7545e16824b86fbd34fa284Author: Adam Jackson <ajax@benzedrine.nwnk.net>Date:   Thu Aug 24 19:52:30 2006 -0400    Enable CPIO in the atimisc driver on ia64 platforms.        This fixes many mach64s, but potentially breaks mach32.  Mach64 is widespread    on ia64 machines but PCI mach32's are quite rare on any platform.commit dfa47b09b5b8d7eeaa3f8ae629ebb3c7efb51511Author: Adam Jackson <ajax@benzedrine.nwnk.net>Date:   Thu Aug 24 19:50:46 2006 -0400    Disable DRI on xpress 200{,m} since we know it's broken.commit 1dcd9b1f75ba2307beb42556aa7f2f945171d5fbAuthor: Adam Jackson <ajax@benzedrine.nwnk.net>Date:   Thu Aug 24 19:42:26 2006 -0400    Disable DRI on RN50 in a more pleasant way.        Rather than publish a non-existant driver name for RN50, just don't bother    initializing it in the first place.commit 64b18e246586b1043a12f381a9d4c834464ed9b5Author: Dave Airlie <airlied@linux.ie>Date:   Fri Aug 25 09:40:04 2006 +1000    update version numbers for releasecommit be229d97e1f0282f2099c7ace7cb257a2738a58fAuthor: Dave Airlie <airlied@linux.ie>Date:   Fri Aug 25 09:39:46 2006 +1000    make colortiling default to on all cardscommit 750ed51bc3d5119e2d2a91a05ae5a971dcd3b88aAuthor: George Sapountzis <gsap7@yahoo.gr>Date:   Mon Aug 7 17:09:02 2006 +0300    [mach64] RENDER support: drop caching DP_WRITE_MASK.        Caching DP_WRITE_MASK causes artifacts, e.g. with evas_xrender_x11_test.commit b2beea2fa8949874d3d57fb9b43fe85cc08a8bffAuthor: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 5 18:36:24 2006 +0300    [mach64] RENDER support: save/restore, cache texture registers.commit 43aaed99950640c3695b3c2b91faabf00c6338a7Author: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 12 22:12:07 2006 +0300    [mach64] RENDER support (bug #6877).        RENDER accleration is disabled by default. It is strongly recommended that the    patch from bug #6772 and the 'exa-damagetrack branch' are merged in the xserver    before enabling RENDER acceleration. To enable RENDER acceleration, add the    following in xorg.conf:        Section "Device"            [...]                Option          "AccelMethod"           "exa"            Option          "RenderAccel"           "true"    EndSectioncommit d113f40fd97408661193895ff3aa82262f100143Author: Dave Airlie <airlied@linux.ie>Date:   Sat Aug 12 17:16:33 2006 +1000    radeon: make some local data static        This is inspired by a patch in XFree86 from David Dawes, it isn't the    same as we have different gamma curves.commit a1da2b5e0a38feae01107676a76bdc5c185ad34bAuthor: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 5 03:18:34 2006 +0300    [mach64] Halve max w/h for hardware-accelerated operations.        EXA hits these limits for some operations on offscreen pixmaps and GTPRO seems    to support the reduced limits.commit a66c8d2f9855caa10142b39f155dd8232b1f2389Author: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 5 03:16:56 2006 +0300    [mach64] EXA: add memcpy-based UTS/DFS.        EXA hits more optimized paths when it does not have to fallback because of    missing UTS/DFS.commit 0fb8503d60cf646c91a806f9ed322f9ceff18baaAuthor: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 5 03:16:15 2006 +0300    [mach64] XAA: minor refactoring of memory manager setup.        - DRI: factor out to separate function and add comments on layout    - no DRI: factor out to separate function, maxScanlines is always equal to      ATIMach64MaxYcommit e203d86643d5d70bf18248712d05b72b79aee705Author: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Aug 5 03:15:42 2006 +0300    [mach64] EXA support.commit dc1e289a611a17090e6dc7ae8a8d3f26d20df4ebAuthor: Samuel Thibault <samuel.thibault@ens-lyon.org>Date:   Sat Jul 29 01:05:35 2006 +0300    Bug #6623: Fix I2C bus Mach64 initialisation failure.commit b377f403e0cd5253402ef3945e5944d1a2763a94Author: Marc Aurele La France <tsi@xfree86.org>Date:   Sat Jul 29 01:03:34 2006 +0300    Fix warnings for Option:1 bit fields.commit 91b8b0ca41ad0b9659f0982a05148cab8558e9d5Author: Marc Aurele La France <tsi@xfree86.org>Date:   Sat Jul 29 01:02:48 2006 +0300    Bug #807: copy throttling.        Fix corruption associated with the engine randomly not waiting for a copy    operation to commit its results.commit 7800407a2db4b7307192bf3b43fffb9c4748b9c0Author: Marc Aurele La France <tsi@xfree86.org>Date:   Sat Jul 29 01:00:56 2006 +0300    Bug #807: invalidate read-back cache.        Fix corruption when scaling 1-pixel wide pixmaps, in which case the first CPU    read of the framebuffer will return stale data.        This bug always manifests itself in mach64 EXA (to be committed soon), with    this patch mach64 EXA passes rendercheck.commit 49b11c540d37152eee4907ab30353ede01e4fb56Author: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Jul 29 00:52:28 2006 +0300    Bug #807: split out TestRegisterCaching block for DP regs.commit 290c5aaa580428bc4748a47e4b9bcc22b90fc8d4Author: George Sapountzis <gsap7@yahoo.gr>Date:   Sat Jul 29 00:51:53 2006 +0300    Bug #807: split out TestRegisterCaching block for DP regs.commit fba8c839b47e2c8d6a6d65950c3431ff5b870aa0Author: Michel D盲nzer <michel@tungstengraphics.com>Date:   Sun Jul 23 18:51:47 2006 +0200    Bug #6819: Fix broken rendering with *8b8g8r8 source formats.        Use the corresponding texture formats with R200 family.        Unfortunately, it looks like the R100 family can't accelerate these source    formats, so we have to fall back for them.commit 60ebe97e2d5d27c2bf4213e201cef74503809cb6Author: Tilman Sauerbeck <tilman@code-monkey.de>Date:   Mon Jul 3 19:37:28 2006 +0200    Move variable declarations before statements to be C89 compatiblecommit 6900ff3c5e574881b1470e2363b200e99ab8bf3eAuthor: Alex Williamson <alex.williamson@hp.com>Date:   Mon Jul 3 13:09:47 2006 +0200    Bug #6872: Disable digital output while reading ROM with Radeon 7000.        This is a workaround for ATI Radeon 7000 erratum F1.commit 920ac1f56fad6f48a30e65836ce054815f270738Author: Michel D盲nzer <michel@tungstengraphics.com>Date:   Mon Jul 3 01:21:18 2006 +0200    Do even less cache flushing when the previous engine mode is known.        OTOH, flush everything when the previous engine mode is unknown, and mark the    engine mode as unknown in a couple more cases.commit dd0dd4bab76907b4b1b6d814f24e73e57d71f587Author: Eric Anholt <anholt@freebsd.org>Date:   Mon Jul 3 01:10:10 2006 +0200    Bug #6762: Remove gratuitous cache flushing.commit 22f165f290f57e638d198f24a37f62fc16b72baeAuthor: Michel D盲nzer <michel@tungstengraphics.com>Date:   Mon Jul 3 00:39:45 2006 +0200    Build and warning fixes for !defined(USE_XAA).commit ea5105230c7c3519730b8d3c250e9ba9a4e4374aAuthor: Michel D盲nzer <michel@tungstengraphics.com>Date:   Mon Jul 3 00:03:22 2006 +0200    Reserve HW cursor memory statically instead of from EXA offscreen.        As a result, quite a bit of code can be removed that attempted to deal with the    EXA offscreen memory disappearing, with varying success. Because the HW cursor    memory is now always immediately after the front buffer and before the back    buffer, this also fixes bug #6808.        As the HW cursor memory is now reserved statically with EXA and XAA, change    RADEONUseHWCursor() to check info->cursor instead of info->cursor_offset, for    which 0 might become valid with HW cursor enabled in the future.