[Intel-gfx] [ANNOUNCE] xf86-video-intel 2.2.99.901

nikosapi me at nikosapi.org
Tue Mar 25 05:07:07 CET 2008


On March 21, 2008 03:53:41 Zhenyu Wang wrote:
> Subject: [ANNOUNCE] xf86-video-intel 2.2.99.901
>
> Here's the first rc for xf86-video-intel 2.3 release. The most noteable
> thing is xvmc branch got merged, and it currently supports XvMC MC hardware
> decode on 915/945/G33 chipsets(i965 hardware decode is still under work,
> and would be in next release though). And we have also fixed several Xv
> relate issues, and with bunch of bug fixes as usual.
>
> Please test for any regressions, and send feedback to us. Thanks!
>
> Changelogs:
>
> Dave Airlie (2):
>       fixup pciaccess version detect
>       quirk: add DVO ivch quirk
>
> Eric Anholt (13):
>       Add detail on different units to intel_idle.
>       Add a little program to output the stepping information for several
> chipsets. Fix SDVO I2C access on Mac Mini in EFI mode.
>       Avoid setting reserved bits in SDVOx on everything 965-class.
>       Add support for brightness control to i915 textured video output.
>       Cleanse DVO drivers of CARD* types.
>       Remove i830+ driver's use of CARD*/INT* types for great justice.
>       Nuke more CARD* types from i810 driver.  Untested.
>       Separate i810 and i830 ring macros out into separate files.
>       Eliminate local variables defined in i830 BEGIN_LP_RING().
>       Add workarounds for 830/845's lack of a8/x8r8g8b8/x8b8g8r8 support.
>       Remove h0h0 UploadToScreen implementation that used pattern blits.
>       Change OUT_RING and similar calls to OUT_BATCH for batchbuffer merge
>
> Hong Liu (7):
>       Fix pixmap offset
>       Bug 10773: fix i8xx pll p2 value in i830_crtc_clock_get()
>       Bug 10584: Mac Mini EDID data assigned to TMDS output
>       Allow non-strict free order for bo_list
>       Fix PLL reference clk debug dump
>       Fix DPLL programming in CRTC mode set
>       Fix rotation mem with kernel memory manager in VT switch
>
> Jesse Barnes (25):
>       Frame buffer compression support on new chipset
>       Program FBC fence offset register
>       Only enable FBC if one pipe is active
>       Fix build warnings on 64 bit
>       Only disable FBC if registers are available
>       Add CACHE_MODE_0 register to dump output
>       Fix DSP*CNTR restoration
>       Remove side effects from VGA debug code
>       Add pipe A force enable quirk for ThinkPad X40
>       Add quirk for DVO channel selection
>       Change FBC idle mode back to default
>       Refactor memory allocation into a separate function
>       Move MTRR fixup to its own routine
>       Account for FBC LL cache in memory reservation
>       Pull chipset detection out of PreInit into its own function
>       Implement new get_property feature for backlight control
>       Enable hardware overlay on 965
>       Add pipe A force enable quirk for Dell Latitude D500
>       Add pipe A force enable quirk for Toshiba Portege R205-S209
>       Add pipe A force quirk for Thinkpad R31
>       Add stdint.h to i830_bios.h
>       Add stdint.h to DVO drivers
>       Add sysfs backlight support for Fujitsu laptops
>       Remove fbc enable/disable messages
>       Add support for new ->get_crtc output callbacks
>
> Julien Cristau (2):
>       Look for src/{modes,parser} in $srcdir, not in objdir
>       Add i8[13]0_ring.h to _SOURCES to fix distcheck
>
> Keith Packard (6):
>       Add intel_idle to measure GPU load ratio.
>       Decode DSPCLK_GATE, dump PIPE*STAT, MI_MODE, MI_DISPLAY_POWER_DOWN,
> MI_ARB_STATE, MI_RDRET_STATE, ECOSKPD 8xx/9xx can handle textures to 2kx2k.
> 965 can do 8kx8k
>       Add defines for 9xx S3 word in 3D_STATE_LOAD_STATE_IMMEDIATE
>       Add i830_transform_is_affine and i830_get_transformed_coordinates_3d.
>       Handle projective transforms on 9xx for Composite.
>
> Kristian Høgsberg (2):
>       Silence warnings about possible uninitialize use of dst_format.
>       Crack down on warnings.
>
> Mark Kettenis (2):
>       Bug #14246: Fix biuld on OpenBSD.
>       Bug #14649: Fix build if pciaccess headers are not with other xorg
> headers.
>
> Matthieu Herrb (1):
>       Makefile.am: nuke RCS Id
>
> Paulo Cesar Pereira de Andrade (2):
>       Make sure symbols used by other modules are public.
>       Compile warning fixes.
>
> Peng Li (1):
>       HP Compaq 6730s has no TV output
>
> Xiang, Haihao (13):
>       enable XvMC for i915
>       fix some bugs
>       fix some bugs
>       Release resource allocated for surface/subpicture
>       fix some bugs
>       fix a bug related to display surface
>       1. The pitches for y/u/v surfaces are 1024bytes aligned.
>       block some signals within critical section to avoid
>       add copyright
>       move i830_allocate_xvmc_buffer out of  #if/#endif
>       lock display when creating context
>       don't initialize XvMC if xv is disabled.
>       only enable xvmc for 915/945/G33
>
> Zhenyu Wang (72):
>       code cleanups
>       remove xvmc attributes
>       explicit say "unsigned int"
>       I missed to remove port attributes wrapper
>       check early if we have slot left for new context or surface
>       disable subpicture initial
>       Add media state definition on i965
>       Create xvmc driver interface
>       comment out i965_xvmc_driver for now
>       fix i915 xvmc devPrivate reference
>       clean up err report
>       Log xvmc driver name
>       Revert to origin xvmc setup steps
>       clean up old i915 xvmc function declare
>       rename xvmc init function names
>       Temp work around ttm allocation for xvmc, which should be fixed
> later. move some definition out of i915 specific header
>       use field name to initialize i915 xvmc driver
>       split xvmc user lib chipset drivers
>       revert xvmc memory allocation hacks
>       add missing license header for new files
>       fix xvmc return values
>       Fix DRI context creation for XvMC
>       try to move some fields to xvmc driver structure
>       xvmc: attempt to move batch buffer code out of i915 private
>       xvmc: take out batch buffer handling code
>       xvmc: move batch buffer init/fini to intel_batchbuffer.c
>       xvmc: oops, should first map batch buffer before init
>       xvmc: move locking code into generic place
>       xvmc: some code clean up
>       xvmc: i915 rename files
>       xvmc: remove unused SAREAPTR macro
>       xvmc: intel_xvmc.c cleanups
>       xvmc: more i915 xvmc cleanups
>       xvmc: remove i915 drm map type
>       xvmc: clean batch buffer finish
>       xvmc: move some i915 context destroy code to generic function
>       xvmc: add xvmc decoder type string
>       xvmc: move batch buffer init/fini code into generic
>       xvmc: add "XvMC" option, default is false.
>       xvmc: move dri context handling to generic code
>       xvmc: handle surface create/destory in generic code
>       xvmc: add disable xvmc build support
>       xvmc: store XvMCContext pointer instead of context_id in internal ctx
> list xvmc: consolidate put surface interfaces
>       xvmc: make XvMC default on
>       xvmc: add info in intel.man
>       xvmc: move drmOpen earlier
>       xvmc: add context id info in debug/trace
>       xvmc: setup driver info for XvMCGetDRInfo request
>       Wrap up chipsets which needs graphics address for status page
>       hardware status page initialization rework
>       Don't crash if SW cursor
>       Fix last commit on i8xx debug p2 value
>       Bug #14440: fix stolen mem size mask on i830M
>       Add DMI info for i830 quirks
>       Fix Lenovo X60 TV quirk
>       Fix last 8XX clock's p2 value commit
>       Don't leak memory if no DMI info is provided by kernel
>       Fix overlay rotation
>       Bug 14591: comment out MI_DISPLAY_POWER_DOWN as for 965GM only
>       Really print adjusted_mode for mode debug
>       No ring, no sync
>       remove extra dri checking
>       disable memory alloc debug
>       Pass default visual in dri context create
>       Always bind memory after agp aperture allocation
>       remove unused init/fini hooks for xvmc driver
>       nuke unneed error_base and event_base
>       bug 15060: Dell Latitude X300 needs pipe A quirk
>       Disable XvMC if DRI is not enabled.
>       Bump version 2.2.99.901
>
> liuhong (1):
>       xf86-video-intel: fix ivch using wrong dvo register
>
> root (3):
>       1. using batch buffer
>       a fix for 945G/945GM
>       1. clean code
>
> git tag: xf86-video-intel-2.2.99.901
>
> http://xorg.freedesktop.org/archive/individual/driver/xf86-video-intel-2.2.
>99.901.tar.bz2 MD5: 0ef9cba695fdfa50002894f863f4a60c 
> xf86-video-intel-2.2.99.901.tar.bz2 SHA1:
> c0d3a3dffabd02f41d2de834c62b5a259302f43e 
> xf86-video-intel-2.2.99.901.tar.bz2
>
> http://xorg.freedesktop.org/archive/individual/driver/xf86-video-intel-2.2.
>99.901.tar.gz MD5: bea6899b89de88873baae08f33b5d70f 
> xf86-video-intel-2.2.99.901.tar.gz SHA1:
> 62a301ef95cf80ba3faf17da4f7c444cd8e3d20a 
> xf86-video-intel-2.2.99.901.tar.gz

Hello,

I'm testing the driver (latest git as of Tue Mar 25 00:03:24 EDT 2008) on my 
Lenovo 3000 V200 laptop with a GM965 (8086:2a02) Integrated Graphics 
Controller.

So far everything seems to work: games, compiz, compiz+xv-video
Using Textured Video with compiz pegs one core of the cpu at 100% and using 
Video Overlay consumes a regular amount of cpu power and looks great (I used 
mplayer -vo xv:port=100 to make it work).

The only issue I've had is the backlight, it worked fine with the Fedora 8 
version (2.1.1-7) of the driver. Now when using my hotkeys the backlight 
flickers but doesn't actually increase (or it'll increase a tiny amount then 
the next time I push the button it'll decrease by that same amount). I can 
still set the level manually through /proc/acpi/video/GFX0/DD02/brightness 
or /sys/class/backlight/acpi_video0/brightness. Applying Jesse Barnes' 
i830-combo-not-legacy.patch doesn't seem to fix the issue.

Another odd thing regarding the backlight is when I play a video using mplayer 
or vlc (not totem though) as soon as the video starts the brightness will be 
reduced to 2 in /sys/class/backlight/acpi_video0/brightness 
and /proc/acpi/video/GFX0/DD02/brightness will stay at whatever value it was 
set at before the video started playing.

Hope this helps,

nick



More information about the Intel-gfx mailing list