From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1765220AbZFLVpt (ORCPT ); Fri, 12 Jun 2009 17:45:49 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S933109AbZFLVpl (ORCPT ); Fri, 12 Jun 2009 17:45:41 -0400 Received: from yw-out-2324.google.com ([74.125.46.31]:49789 "EHLO yw-out-2324.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1757791AbZFLVpj convert rfc822-to-8bit (ORCPT ); Fri, 12 Jun 2009 17:45:39 -0400 DomainKey-Signature: a=rsa-sha1; c=nofws; d=gmail.com; s=gamma; h=mime-version:in-reply-to:references:date:message-id:subject:from:to :cc:content-type:content-transfer-encoding; b=aQmnTcn4SOFB6lTjEIXa+MFToAYAd4K20dzmPgg4PysfTu1RdzXl1MAyqES79Tqt8P FhwgaPpAXEzMostue0F62uPSMcaiWEQvZZOcbHwfvNkb0bOVc/cnW72Q9OrEX2weKEm3 Nc3UM3Y6/zABID+3qqbK9nO42AB6Joy+1o51Q= MIME-Version: 1.0 In-Reply-To: References: Date: Sat, 13 Jun 2009 07:45:40 +1000 Message-ID: <21d7e9970906121445j503671c5yedbd9cecf20f0643@mail.gmail.com> Subject: Re: [git pull] drm v2.6.31 merge (part 1) From: Dave Airlie To: Dave Airlie Cc: torvalds@linux-foundation.org, linux-kernel@vger.kernel.org Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8BIT Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Fri, Jun 12, 2009 at 4:28 PM, Dave Airlie wrote: > > Hi Linus, > > Please pull the 'drm-linus' branch from > ssh://master.kernel.org/pub/scm/linux/kernel/git/airlied/drm-2.6.git drm-linus > Did I do something wrong again? or are you just holding off on this, didn't see it hit the last bunch of merges. Dave. > This contains the Intel tree merge (merged properly I haven't rebased or > touched it), which contains numerous GEM bugfixes + support for a new > chipset. AMD patches for new r600 chipset support. A more flexible drm > debugging system to decrease the firehose effect enabling drm debugging > has, it also contains some paving the way patches for part 2 of the merge. > > It also contains one AGP patch for supporting those new chips, and a PNP > patch to add a new interface that intel kms relies on now, all the > signoffs for the pnp code should be correct. > > Part 2 will contains an initial radeon KMS driver and the TTM memory > manager, its quite large so I don't want to include it all in this pull. > The initial radeon KMS code enable switch will hide under staging for now > for one driver release while we stabilise it in-tree, its not in a bad > state but its a lot of new code and we'd hate for anyone to fall over it > my accident. Its quite well separated from the old radeon code so > shouldn't fall over too much. I'll send the part 2 pull early next week. > > Dave. > >  drivers/char/agp/intel-agp.c           |   16 +- >  drivers/gpu/drm/drm_bufs.c             |    3 +- >  drivers/gpu/drm/drm_edid.c             |   74 +---- >  drivers/gpu/drm/drm_gem.c              |    2 +- >  drivers/gpu/drm/drm_hashtab.c          |    4 + >  drivers/gpu/drm/drm_mm.c               |  165 +++++++-- >  drivers/gpu/drm/drm_modes.c            |   18 +- >  drivers/gpu/drm/drm_stub.c             |   15 + >  drivers/gpu/drm/i915/i915_dma.c        |   67 ++-- >  drivers/gpu/drm/i915/i915_drv.h        |   48 ++- >  drivers/gpu/drm/i915/i915_gem.c        |  156 ++++++--- >  drivers/gpu/drm/i915/i915_gem_tiling.c |  152 ++++++++ >  drivers/gpu/drm/i915/i915_irq.c        |  190 +++++++++- >  drivers/gpu/drm/i915/i915_reg.h        |  616 ++++++++++++++++++++++++++++++- >  drivers/gpu/drm/i915/i915_suspend.c    |   20 + >  drivers/gpu/drm/i915/intel_bios.c      |   86 +++++- >  drivers/gpu/drm/i915/intel_bios.h      |  101 +++++- >  drivers/gpu/drm/i915/intel_crt.c       |   76 ++++- >  drivers/gpu/drm/i915/intel_display.c   |  645 ++++++++++++++++++++++++++++++-- >  drivers/gpu/drm/i915/intel_fb.c        |   26 +- >  drivers/gpu/drm/i915/intel_hdmi.c      |   33 ++- >  drivers/gpu/drm/i915/intel_lvds.c      |  151 ++++++-- >  drivers/gpu/drm/i915/intel_sdvo.c      |  110 ++++-- >  drivers/gpu/drm/i915/intel_tv.c        |    3 + >  drivers/gpu/drm/radeon/r600_cp.c       |   42 ++- >  drivers/gpu/drm/radeon/radeon_cp.c     |    2 +- >  drivers/gpu/drm/radeon/radeon_drv.h    |    1 + >  drivers/gpu/drm/via/via_dmablit.c      |    6 +- >  drivers/pnp/resource.c                 |   18 + >  include/drm/drmP.h                     |  126 ++++--- >  include/drm/drm_hashtab.h              |    2 + >  include/drm/drm_mm.h                   |   90 +++++ >  include/drm/drm_pciids.h               |    9 + >  include/linux/pnp.h                    |    2 + >  34 files changed, 2677 insertions(+), 398 deletions(-) >  create mode 100644 include/drm/drm_mm.h > > commit 3c24475c1e4e8d10e50df161d8c4f1d382997a7c > Author: Jerome Glisse > Date:   Wed Apr 8 18:34:28 2009 +0200 > >    drm: include kernel list header file in hashtab header > >    Signed-off-by: Dave Airlie > > commit f2cb5d86e1af175a9b210241800f03a447f92621 > Author: Jerome Glisse > Date:   Wed Apr 8 17:16:24 2009 +0200 > >    drm: Export hash table functionality. > >    add exports so TTM module can use these functions. > >    Signed-off-by: Thomas Hellstrom >    Signed-off-by: Dave Airlie > > commit 249d6048ca98b5452105b0824abac1275661b8e3 > Author: Jerome Glisse > Date:   Wed Apr 8 17:11:16 2009 +0200 > >    drm: Split out the mm declarations in a separate header. Add atomic operations. > >    this is a TTM preparation patch, it rearranges the mm and >    add operations needed to do mm operations in atomic context. > >    Signed-off-by: Thomas Hellstrom >    Signed-off-by: Dave Airlie > > commit 715cbb05c935e8a4306a730d14a72d5af881523e > Author: Alex Deucher > Date:   Fri Jun 12 15:55:44 2009 +1000 > >    drm/radeon: add support for RV790. > >    This adds the PCI IDs for the rv790 which are equiv to the rv770. > >    Signed-off-by: Dave Airlie > > commit 2a71ebcd85bcc4d6607f577f23a491f796c30e82 > Author: Alex Deucher > Date:   Fri Jun 12 15:53:10 2009 +1000 > >    drm/radeon: add rv740 drm support. > >    This adds drm support for the RV740 family of chips to the r600 support code. > >    Signed-off-by: Dave Airlie > > commit fbe0efb869efde8d847ede3a925230ef88910086 > Author: Kristian Høgsberg > Date:   Tue Jun 9 01:50:41 2009 +1000 > >    drm_calloc_large: check right size, check integer overflow, use GFP_ZERO > >    Previously we would check size instead of size * nmemb, and so would >    never hit the vmalloc path.  Also add integer overflow check as in kcalloc, >    and allocate GFP_ZERO pages instead of memset()ing them. > >    Signed-off-by: Kristian Høgsberg >    Signed-off-by: Dave Airlie > > commit 61f11699e7a92d932b31ded3715ad4f70eb26ef2 > Author: Keith Packard > Date:   Sat May 30 20:42:27 2009 -0700 > >    drm: Eliminate magic I2C frobbing when reading EDID > >    This code depends on the underlying I2C adapter using the bit-banging algo, >    which may not be the case. If specific encoders require this mechanism, they >    should build a custom I2C algo that implements this workaround, rather than >    having it in the general path. > >    Signed-off-by: Keith Packard >    Signed-off-by: Dave Airlie > > commit 7ff145593d808a371924652c8d6a15fb75ce2250 > Author: Dave Airlie > Date:   Wed Apr 22 18:52:14 2009 +1000 > >    drm/i915: duplicate desired mode for use by fbcon. > >    duplicate the mode into fbcon storage, so when we free modes later >    we don't just lose this. > >    Signed-off-by: Dave Airlie > > commit c5c07550d41abe86d109430e718f2007113031f8 > Author: Figo.zhang > Date:   Sat Jun 6 18:26:26 2009 +0800 > >    drm/via: vfree() no need checking before calling it > >    vfree() does it's own NULL checking, no need for explicit check before >    calling it. > >    Signed-off-by: Figo.zhang >    Signed-off-by: Dave Airlie > > commit be25ed9c5cc06e1d17aa97e41daf88f0b46143e6 > Author: yakui_zhao > Date:   Tue Jun 2 14:13:55 2009 +0800 > >    drm: Replace DRM_DEBUG with DRM_DEBUG_DRIVER in i915 driver > >    Replace the DRM_DEBUG with the DRM_DEBUG_DRIVER to print the debug info >    in i915 driver. > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Dave Airlie > > commit f053185948a1bd16329433f5371809765086c1ec > Author: yakui_zhao > Date:   Tue Jun 2 14:12:47 2009 +0800 > >    drm: Replace DRM_DEBUG with DRM_DEBUG_MODE in drm_mode > >    Replace the DRM_DEBUG with DRM_DEBUG_MODE macro to print the info in drm_mode. > >    airlied:- fixed up to remove a conflicting #define >    Signed-off-by: Zhao Yakui >    Signed-off-by: Dave Airlie > > commit 342dc382c451f75ea202a65e6f529bdff6d184cd > Author: yakui_zhao > Date:   Tue Jun 2 14:12:00 2009 +0800 > >    drm/i915: Replace DRM_DEBUG with DRM_DEBUG_KMS in intel_sdvo > >    Use the DRM_DEBUG_KMS/DEBUG_LOG_KMS to print the debug info for >    SDVO device. > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Dave Airlie > > commit 7fb85bfb54a64e9dd82ee4a79022c38ab58f21a0 > Author: yakui_zhao > Date:   Tue Jun 2 14:10:49 2009 +0800 > >    drm/i915: replace DRM_DEBUG with DRM_DEBUG_KMS in intel_lvds > >    Use the DRM_DEBUG_KMS macro definition to print the debug info for >    the LVDS. > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Dave Airlie > > commit 4fefcb27050b98c97b1c32bc710fc2f874449dee > Author: yakui_zhao > Date:   Tue Jun 2 14:09:47 2009 +0800 > >    drm: add separate drm debugging levels > >    Now all the DRM debug info will be reported if the boot option of >    "drm.debug=1" is added. Sometimes it is inconvenient to get the debug >    info in KMS mode. We will get too much unrelated info. > >    This will separate several DRM debug levels and the debug level can be used >    to print the different debug info. And the debug level is controlled by the >    module parameter of drm.debug > >    In this patch it is divided into four debug levels; >                drm_core, drm_driver, drm_kms, drm_mode. > >    At the same time we can get the different debug info by changing the debug >    level. This can be done by adding the module parameter. Of course it can >    be changed through the /sys/module/drm/parameters/debug after the system is >    booted. > >    Four debug macro definitions are provided. >        DRM_DEBUG(fmt, args...) >        DRM_DEBUG_DRIVER(prefix, fmt, args...) >        DRM_DEBUG_KMS(prefix, fmt, args...) >        DRM_DEBUG_MODE(prefix, fmt, args...) > >    When the boot option of "drm.debug=4" is added, it will print the debug info >    using DRM_DEBUG_KMS macro definition. >    When the boot option of "drm.debug=6" is added, it will print the debug info >    using DRM_DEBUG_KMS/DRM_DEBUG_DRIVER. > >    Sometimes we expect to print the value of an array. >    For example: SDVO command, >    In such case the following four DRM debug macro definitions are added: >        DRM_LOG(fmt, args...) >        DRM_LOG_DRIVER(fmt, args...) >        DRM_LOG_KMS(fmt, args...) >        DRM_LOG_MODE(fmt, args...) > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Dave Airlie > > commit df4f7fe7bd516b3833e25c692c3970e22038a6ca > Author: Dave Airlie > Date:   Thu Jun 11 16:16:10 2009 +1000 > >    radeon: remove _DRM_DRIVER from the preadded sarea map > >    This shouldn't be there and is what broke r600 late in the 2.6.30 >    release cycle with Ben's patch. > >    Signed-off-by: Dave Airlie > > commit 2ff2e8a3e1898e692b604424c384f134009dea80 > Author: Ben Skeggs > Date:   Tue May 26 10:35:52 2009 +1000 > >    drm: don't associate _DRM_DRIVER maps with a master > >    A driver will use the _DRM_DRIVER map flag to indicate that it wants >    to be responsible for removing the map itself, bypassing the DRM's >    automagic cleanup code. > >    Since the multi-master changes this has been broken, resulting in some >    drivers having their registers unmapped before it's finished with them. > >    Signed-off-by: Ben Skeggs >    Signed-off-by: Dave Airlie > > commit b798b1fe3b6436275ad1b517a6823d55e3b75c22 > Author: Robert P. J. Day > Date:   Wed Jun 10 12:43:49 2009 -0700 > >    drm: simplify kcalloc() call to kzalloc(). > >    Calls to kcalloc() for a single element can be simplified to calls to >    kzalloc(). > >    Signed-off-by: Robert P. J. Day >    Cc: Dave Airlie >    Cc: Eric Anholt >    Signed-off-by: Andrew Morton >    Signed-off-by: Dave Airlie > > commit 1bcbf3948876e31a8ece28597dec447611ad9c8b > Author: Pavel Roskin > Date:   Wed Jun 10 12:43:48 2009 -0700 > >    intelfb: fix spelling of "CLOCK" > >    Signed-off-by: Pavel Roskin >    Cc: Eric Anholt >    Cc: Dave Airlie >    Signed-off-by: Andrew Morton >    Signed-off-by: Dave Airlie > > commit dcae3626d031fe6296b1e96a16f986193a41f840 > Author: Roel Kluin > Date:   Wed Jun 10 12:43:48 2009 -0700 > >    drm: fix LOCK_TEST_WITH_RETURN macro > >    When this macro isn't called with 'file_priv' this will result in a build >    failure. > >    Signed-off-by: Roel Kluin >    Signed-off-by: Andrew Morton >    Signed-off-by: Dave Airlie > > commit 03d6069912babc07a3da20e715dd6a5dc8f0f867 > Author: Keith Packard > Date:   Fri Jun 5 18:19:56 2009 -0700 > >    drm/i915: Hook connector to encoder during load detection (fixes tv/vga detect) > >    With the DRM-driven DPMS code, encoders are considered idle unless a >    connector is hooked to them, so mode setting is skipped. This makes load >    detection fail as none of the hardware is enabled. > >    Signed-off-by: Keith Packard >    Signed-off-by: Eric Anholt > > commit 2939e1f5331455d17a4a704dd6210e1474002545 > Author: Chris Wilson > Date:   Sat Jun 6 09:46:03 2009 +0100 > >    drm/i915: NOMEM->NOSPC > >    To differentiate between encountering an out-of-memory error with running >    out of space in the aperture, use ENOSPC for the later. > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit 21d509e339565c82887733c02465bb7f5866c8f5 > Author: Chris Wilson > Date:   Sat Jun 6 09:46:02 2009 +0100 > >    drm/i915: use I915_GEM_GPU_DOMAINS > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit b1ce786cb85280490ca3c29a62ddf8608826b414 > Author: Chris Wilson > Date:   Sat Jun 6 09:46:00 2009 +0100 > >    drm/i915: no need to hold mutex for object lookup > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit 1f803ee5cea67d2387aeedb4b07e645a743729de > Author: Chris Wilson > Date:   Sat Jun 6 09:45:59 2009 +0100 > >    drm/i915: Call drm_vblank_post_modeset() on error paths. > >    Ensure that the drm_vblank_pre_modeset() is always balanced by >    drm_vblank_post_modeset() within intel_crtc_mode_set(). > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit 5f26a2c7ad6eba97141e8372f3def282f934b169 > Author: Chris Wilson > Date:   Sat Jun 6 09:45:58 2009 +0100 > >    drm/i915: OR in the COMMAND read domain for the batch buffer. > >    The batch buffer may be shared with another read buffer, so we should not >    ignore any previously set domains, but just or in the command domain (and >    check that the buffer is not writable). > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit 83d60795157c83389e6aaa0532d5e19afa976a24 > Author: Chris Wilson > Date:   Sat Jun 6 09:45:57 2009 +0100 > >    drm/i915: Sanity check execbuffer arguments before touching state. > >    By sending a broken execbuffer (its length was not suitably aligned) I >    triggered an operation upon a freed object. The invalid alignment was >    discovered after updating the write_domain on the object but before the >    object was placed on the active queue. So during the unwind process >    following the error, the now freed object attempts to flush its >    non-existent, but outstanding, GPU writes causing this use-after-free. > >    [drm:i915_dispatch_gem_execbuffer] *ERROR* alignment >    [drm:i915_gem_execbuffer] *ERROR* dispatch failed -22 >    WARNING: at lib/kref.c:43 warn_slowpath_null+0x10/0x15() >    Modules linked in: >    Pid: 4552, comm: lt-csi-drm Not tainted 2.6.30-rc6 #423 >    Call Trace: >     [] warn_slowpath_fmt+0x57/0x6d >     [] ? get_pageblock_migratetype+0x18/0x1e >     [] ? free_hot_page+0xa/0xc >     [] ? __free_pages+0x16/0x1f >     [] ? shmem_truncate_range+0x63e/0x656 >     [] ? slob_page_alloc+0x146/0x1c8 >     [] warn_slowpath_null+0x10/0x15 >     [] kref_get+0x1b/0x21 >     [] i915_gem_object_move_to_active+0x1f/0x56 >     [] i915_add_request+0x156/0x19a >     [] i915_gem_object_flush_gpu_write_domain+0x28/0x3f >     [] i915_gem_object_unbind+0x4a/0x124 >     [] i915_gem_free_object+0x33/0x9b >     [] drm_gem_object_free+0x28/0x4a >     [] ? drm_gem_object_free+0x0/0x4a >     [] kref_put+0x38/0x41 >     [] drm_gem_object_unreference+0x11/0x13 >     [] drm_gem_object_handle_unreference+0x1e/0x21 >     [] drm_gem_object_release_handle+0xa/0xe >     [] idr_for_each+0x5f/0x98 >     [] ? drm_gem_object_release_handle+0x0/0xe >     [] drm_gem_release+0x22/0x34 >     [] drm_release+0x1e8/0x3c4 >     [] __fput+0xaf/0x146 >     [] fput+0x12/0x14 >     [] filp_close+0x48/0x52 >     [] put_files_struct+0x57/0x9b >     [] exit_files+0x1e/0x20 >     [] do_exit+0x16d/0x511 >     [] ? __schedule+0x3d4/0x3e5 >     [] ? handle_irq+0xd/0x69 >     [] do_group_exit+0x4d/0x73 >     [] sys_exit_group+0x13/0x17 >     [] sysenter_do_call+0x12/0x2b > >    Signed-off-by: Chris Wilson >    Signed-off-by: Eric Anholt > > commit fa0864b26b4bfa1dd4bb78eeffbc1f398cb56425 > Author: Michael Cousin > Date:   Fri Jun 5 21:16:22 2009 +0200 > >    drm/i915: Skip lvds with Aopen i945GTt-VFA > >    Signed-off-by: Michael Cousin >    Signed-off-by: Eric Anholt > > commit 80a538e49daddbf3bf783f3464e91bd3181957b2 > Author: Zhenyu Wang > Date:   Mon Jun 8 14:40:20 2009 +0800 > >    drm/i915: Enable probe on new chipset > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit 036a4a7d9272582fc7370359515d807393e2f728 > Author: Zhenyu Wang > Date:   Mon Jun 8 14:40:19 2009 +0800 > >    drm/i915: handle interrupt on new chipset > >    Update interrupt handling methods for IGDNG with new registers >    for display and graphics interrupt functions. As we won't use >    irq-based vblank sync in dri2, so display interrupt on new chip >    will be used for hotplug only in future. > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit d765898970f35acef960581f678b9da9d5c779fa > Author: Jesse Barnes > Date:   Fri Jun 5 14:41:29 2009 +0000 > >    drm/i915: enable MCHBAR if needed > >    Using the new PNP resource checking code, this patch allows the i915 >    driver to allocate MCHBAR space if needed and use the BAR to determine >    current memory settings. > >    [apw@canonical.com: moved to the new generic PNP resource interface] >    Signed-off-by: Jesse Barnes >    Signed-off-by: Andy Whitcroft >    Signed-off-by: Eric Anholt > >    failure to update-index after git-am --reject to hand-apply > >    Signed-off-by: Eric Anholt > > commit 1b8e69662e1a086878bf930a6042daf7f8a076cc > Author: Bjorn Helgaas > Date:   Fri Jun 5 14:37:23 2009 +0000 > >    pnp: add PNP resource range checking function > >    Add a PNP resource range check function, indicating whether a resource >    has been assigned to any device. > >    Signed-off-by: Bjorn Helgaas >    [apw@canonical.com: fixed up exports et al] >    Signed-off-by: Andy Whitcroft >    Signed-off-by: Eric Anholt > > commit b66d18ddb16603d1e1ec39cb2ff3abf3fd212180 > Author: Jesse Barnes > Date:   Fri May 15 14:11:48 2009 -0700 > >    drm/i915: avoid non-atomic sysrq execution > >    The sysrq functions are executed in hardirq context, so we shouldn't be >    calling sleeping functions from them, like mutex_locks or memory >    allocations. > >    Fix up the i915 sysrq handler to avoid this. > >    Signed-off-by: Jesse Barnes >    Signed-off-by: Eric Anholt > > commit 42c2798b35b95c471877133e19ccc3cab00e9b65 > Author: Jesse Barnes > Date:   Tue May 5 13:13:16 2009 -0700 > >    drm/i915: apply G45 vblank count code to all G4x chips and fix max_frame_count > >    All G4x and newer chips use the new style frame count register, with a >    full 32 bit frame count.  Update the code to reflect this. > >    Signed-off-by: Jesse Barnes >    Signed-off-by: Eric Anholt > > commit 70aa96ca2d8d938fc036ef8fd189b0151f4fc3ba > Author: Jarod Wilson > Date:   Wed May 27 17:20:39 2009 -0400 > >    drm/i915: add ignore lvds quirk info for AOpen Mini PC > >    Fix a FIXME in the intel LVDS bring-up code, adding the appropriate >    blacklist entry for the AOpen Mini PC, courtesy of a dmidecode >    dump from Florian Demmer. > >    Signed-off-by: Jarod Wilson >    CC: Florian Demmer >    Signed-off-by: Eric Anholt > > commit 714605e4a05787c51a5ac36c926d2169cfdfbfba > Author: yakui_zhao > Date:   Sun May 31 17:18:07 2009 +0800 > >    drm/i915: Initialize the SDVO device based on the sdvo info parsed from VBT > >    http://bugs.freedesktop.org/show_bug.cgi?id=20429 > >    Signed-off-by: Zhao Yakui >    [anholt: Massive cleanup of the slave addr function] >    Signed-off-by: Eric Anholt > > commit 9b9d172d06b0f2d51cc9431e2c6c3055f0cf10ef > Author: yakui_zhao > Date:   Sun May 31 17:17:17 2009 +0800 > >    drm/i915: parse VBT general definition block to get the SDVO device info > >    The general definition block contains the child device tables, which include >    the SDVO device info. For example: device slave address, device dvo port, >    device type. > >    We will get the info of SDVO device by parsing the general definition blocks. >    Only when a valid slave address is found, it is regarded as the SDVO device. >    And the info of DVO port and slave address is recorded. > >    http://bugs.freedesktop.org/show_bug.cgi?id=20429 > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Eric Anholt > > commit 59a036cfbd29aadf40d2b754cfebee2a96268752 > Author: yakui_zhao > Date:   Sun May 31 17:16:22 2009 +0800 > >    drm/i915: Add the structure of child_device_config in video BIOS tables. > >    Signed-off-by: Zhao Yakui >    Signed-off-by: Eric Anholt > > commit 040d87f15a0129242463d3ed7c48381505f596e2 > Author: Keith Packard > Date:   Sat May 30 20:42:33 2009 -0700 > >    drm/i915: Add Display Port register defines > >    This adds the register definitions for the display port enable register >    along with those for the GMCH and Link M/N ratios required to drive display >    port outputs. > >    Signed-off-by: Keith Packard >    Signed-off-by: Eric Anholt > > commit 2245fda810f870dce9b030e6aa604320abba53a5 > Author: Keith Packard > Date:   Sat May 30 20:42:29 2009 -0700 > >    drm/i915: Don't trim cursor addresses to 11 bits > >    We can safely assume that cursor addresses will not extend beyond the >    addressable screen dimensions; setting the additional bits is harmless in >    any case. > >    Signed-off-by: Keith Packard >    Signed-off-by: Eric Anholt > > commit cb66c692d1ae257f32dc7f6085cf9cb9f2f6bab8 > Author: Ma Ling > Date:   Sun May 31 16:58:32 2009 +0800 > >    drm/i915: Set correct TV detection voltage level override values > >    We detect TV connect status by setting DAC voltage level override >    values as 0.7 voltage for DAC_A/B/C. The corresponding 2-bits shold be 0x2, >    In order correctly to set last bit as 0, at first we must clean it. > >    It fixed freedesktop.org bug #21204 > >    Signed-off-by: Ma Ling >    Signed-off-by: Eric Anholt > > commit 541998a18b72d2cac48b3369fa4540116ff3f0a8 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:44 2009 +0800 > >    drm/i915: Add LVDS support for IGDNG > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit 30ad48b7334a2eb2edf22f6c91f7b3f22a22a837 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:43 2009 +0800 > >    drm/i915: Add HDMI support on IGDNG > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit 2c07245fb8f7f0a282282e5a9747e46defdb2cc7 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:42 2009 +0800 > >    drm/i915: enable kernel modesetting on IGDNG > >    This adds kernel mode setting on IGDNG with VGA output support. >    Note that suspend/resume doesn't work yet. > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit 2cce0d8740f0d1454d012401257d96c513ce358f > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:41 2009 +0800 > >    drm/i915: Disable tiling on IGDNG for now > >    Swizzle bit detection not working right on it. > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit e170b030dcd6aed11dde2d124c09991ec771f529 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:40 2009 +0800 > >    drm/i915: Disable opregion on IGDNG for now > >    Disable OpRegion support for now until verified on new chipsets. > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit b9055052d3e0388b4a5e8c3e0bbab665c5996f50 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:38 2009 +0800 > >    drm/i915: Add new chipset register definitions > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit 280da227c870a50f669de0c8d46bfb2c62da9995 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:37 2009 +0800 > >    drm/i915: Add chipset/feature defines for for new chipsets > >    Signed-off-by: Zhenyu Wang >    [anholt: dropped drm_pciids.h hunk to avoid loading an incomplete driver] >    Signed-off-by: Eric Anholt > > commit 32cb055b57eab803ea82b76dc913b0378e5af145 > Author: Zhenyu Wang > Date:   Fri Jun 5 15:38:36 2009 +0800 > >    agp/intel: Add support for new chipsets > >    Both desktop and mobile versions are added. > >    Signed-off-by: Zhenyu Wang >    Signed-off-by: Eric Anholt > > commit b962442e46a9340bdbc6711982c59ff0cc2b5afb > Author: Eric Anholt > Date:   Wed Jun 3 07:27:35 2009 +0000 > >    drm/i915: Change GEM throttling to be 20ms like the comment says. > >    keithp didn't like the original 20ms plan because a cooperative client could >    be starved by an uncooperative client.  There may even have been problems >    with cooperative clients versus cooperative clients.  So keithp changed >    throttle to just wait for the second to last seqno emitted by that client. >    It worked well, until we started getting more round-trips to the server >    due to DRI2 -- the server throttles in BlockHandler, and so if you did more >    than one round trip after finishing your frame, you'd end up unintentionally >    syncing to the swap. > >    Fix this by keeping track of the client's requests, so the client can wait >    when it has an outstanding request over 20ms old.  This should have >    non-starving behavior, good behavior in the presence of restarts, and less >    waiting.  Improves high-settings openarena performance on my GM45 by 50%. > >    Signed-off-by: Eric Anholt >    Reviewed-by: Jesse Barnes > > commit 1fd1c624362819ecc36db2458c6a972c48ae92d6 > Author: Eric Anholt > Date:   Wed Jun 3 07:26:58 2009 +0000 > >    drm/i915: Save/restore cursor state on suspend/resume. > >    This may fix cursor corruption in X on resume, which would persist until >    the cursor was hidden and then shown again. > >    V2: Also include the cursor control regs. > >    Signed-off-by: Eric Anholt >    Reviewed-by: Jesse Barnes > > commit 0e7ddf7eeeef5aea85412120539ab5369577faeb > Author: Eric Anholt > Date:   Thu Jun 4 11:18:14 2009 +0000 > >    drm/i915: Remove a bad BUG_ON in the fence management code. > >    This could be triggered by a gtt mapping fault on 965 that decides to >    remove the fence from another object that happens to be active currently. >    Since the other object doesn't rely on the fence reg for its execution, we >    don't wait for it to finish.  We'll soon be not waiting on 915 most of the >    time as well, so just drop the BUG_ON. > >    Signed-off-by: Eric Anholt