Michel Dänzer
a8d6ebdf93
EXA: Defer to FillRegionTiled in Composite when possible.
...
Committed separately as this case is hard to hit and has only been tested
lightly.
2007-04-29 23:47:53 +02:00
Michel Dänzer
81b055605c
EXA: Composite improvements.
...
* Defer to simpler hooks in more cases (inspired by XAA behaviour).
* Move damage tracking from lower to higher level functions.
* Always migrate for fallbacks.
2007-04-29 23:47:43 +02:00
Michel Dänzer
ce317a5b76
EXA: Glyphs improvements.
...
* Don't waste effort on invisible glyphs.
* Add damage tracking where necessary.
* Always migrate for fallbacks.
2007-04-29 23:47:16 +02:00
Michel Dänzer
0c8905ebc9
EXA: PolyFillRect improvements.
...
* Convert rects to region and use it for damage tracking.
* When possible, defer to exaFillRegion{Solid,Tiled} using converted region.
* Always migrate for fallbacks.
* Move damage tracking out of ExaCheckPolyFillRect.
2007-04-29 23:47:08 +02:00
Michel Dänzer
567f18a09b
EXA: FillRegion{Solid,Tiled} improvements.
...
* Support planemasks, different ALUs and arbitrary tile origin.
* Leave damage tracking and non-trivial fallbacks to callers.
* Always migrate for fallbacks.
This is in preparation for using these from more other functions.
2007-04-29 23:46:49 +02:00
Michel Dänzer
e869573b52
EXA: exaAssertNotDirty improvements.
...
* Return early if the valid region is empty or the pixmap is pinned.
* Fix loop for several cliprects.
2007-04-29 23:45:48 +02:00
Michel Dänzer
d3f8667341
EXA: Fix exaEnableDisableFBAccess for nested disables and enables.
2007-04-29 23:44:27 +02:00
Michel Dänzer
5e4b3232da
Fix fbCompositeTrans_0888xnx0888 build for wfb on big endian.
2007-04-29 23:38:22 +02:00
Michel Dänzer
2866e0bac9
Fix a couple of picture repeat fields incorrectly compared to RepeatNormal.
2007-04-29 23:38:13 +02:00
Soren Sandmann Pedersen
78a2045535
Pixman merging
...
More msvc++ porting
2007-04-27 15:20:24 -04:00
Soren Sandmann Pedersen
6c4f1826bf
Bug fix in fbCompositeIn_nx8x8888
...
Make sure both halves of the dst word is set to zero when the masks
are both 0.
2007-04-27 08:13:45 -04:00
Adam Jackson
ae04f2cb0a
Fix the 'relink' target for kdrive servers.
2007-04-26 15:28:04 -04:00
Adam Jackson
6c8152d6ee
Remove old edid_modes.c, it lives in hw/xfree86/modes/ now.
2007-04-26 14:59:04 -04:00
Soren Sandmann Pedersen
2208c6087d
Change expand_alpha_rev to expand_alpha in mmxSaturateU
2007-04-26 14:40:30 -04:00
Soren Sandmann Pedersen
a300ef84ce
Fix typo in previous commit
2007-04-26 14:37:53 -04:00
Soren Sandmann Pedersen
0ebe48be59
Pixman merge
...
- Changes to support MS Visual C++
- use inline instead of __inline__
- Fix rounding errors (Billy Biggs, from xserver via pixman)
2007-04-26 14:36:32 -04:00
Soren Sandmann Pedersen
4fe918b385
Fix bug 8871 - scrolling corruption with a compositing manager
...
Call miHandleExposures() in CopyArea/CopyPlane explicitly in cw to
generate GraphicsExposes correctly.
2007-04-26 12:41:01 -04:00
Soren Sandmann Pedersen
0ff7c94fcf
Pixman merge
...
Make use of fbCompositeSrcAdd_8888x8x8mmx
2007-04-26 10:56:02 -04:00
Soren Sandmann Pedersen
701ccb4a22
Pixman merge
...
- Remove stray default label
- Integrate new MMX ops SolidMaskSrc_nx8x8888mmx, In_8x8mmx, and
In_nx8x8mmx
- Formatting changes to reduce diff noise
2007-04-26 10:49:06 -04:00
Soren Sandmann Pedersen
a54ef54db1
Pixman merge
...
Make sure fbCompositeSrc_x888x8x8888mmx and fbCompositeSrc_8888x8x8888mmx
are used when possible.
2007-04-26 10:24:25 -04:00
Soren Sandmann Pedersen
3571b8e65b
More pixman merging
...
- Move some code around to minimize diff noise
- Add mmx ops from pixman that never made it into X
- Merge Jeff Muizelaar's bugfixes to fbCompositeSrc_8888x8x8888mmx and
fbCompositeSrc_x888x8x8888mmx.
2007-04-26 10:06:47 -04:00
Peter Hutterer
c0346e57e6
Require inputproto 1.4.2.
...
Requirement was introduced with c7e2ba0c9b
.
2007-04-26 12:03:28 +09:30
Soren Sandmann Pedersen
67347739b0
Don't treat convolution filters as transformations.
...
Some rearrangement of code to get it closer to pixman.
2007-04-25 22:22:22 -04:00
Soren Sandmann Pedersen
c056ce95d8
Port MSVC++ CPU detection code from pixman. (Vladimir Vukicevic).
2007-04-25 22:22:22 -04:00
Soren Sandmann Pedersen
c19ece1d8c
Integrate optimization from xserver from David Reveman where repeats
...
get handled by fbFetchTransformed() rather than in the region walking
code.
2007-04-25 22:22:22 -04:00
Soren Sandmann Pedersen
48c73dfc36
Add function fbCompositeSrcAdd_8888x8x8(), and fix a bug where
...
srcRepeat = FALSE would be set in the wrong place.
2007-04-25 22:22:22 -04:00
Soren Sandmann Pedersen
66ba3d758a
Various fixes from xserver via pixman (Billy Biggs)
2007-04-25 22:22:22 -04:00
Adam Jackson
c09e68ce30
Paper over a crash at exit during GLX teardown.
2007-04-25 16:46:26 -04:00
Adam Jackson
9c80eda826
Disable RANDR's fake Xinerama protocol when there's more than one screen.
...
... in the protocol sense. Xinerama doesn't have any provision for more
than one protocol screen each with its own geometry.
Red Hat bug #231257 .
2007-04-25 16:35:04 -04:00
Adam Jackson
d322608dc9
Fix a buffer overrun on machines with excessively large PCI busses.
...
Formerly we sized an array with a compile time constant, then initialized
its size to the same constant, but the Linux PCI init code would increase
that "constant". So if you happened to have more than 128 PCI devices,
you'd happily scribble into whatever variables happened to be in .bss
after that array.
Only really fixed for Linux atm. Other OSes will simply (still) fail to
work on video devices above the 128th PCI device.
2007-04-25 16:29:48 -04:00
Aaron Plattner
3ba1e8ab6d
Include xf86Rename.h in xf86RandR12.h.
2007-04-24 17:20:14 -07:00
Soren Sandmann Pedersen
0a2fe443d2
Use READ/WRITE macros for new functions introduced in previous commits.
2007-04-24 19:02:44 -04:00
Soren Sandmann Pedersen
7e16da7b78
Remove #if 0'ed blocks
2007-04-24 18:15:34 -04:00
Brian
18252a515d
bump release date to reflect input code updates
2007-04-24 14:10:09 -06:00
Soren Sandmann Pedersen
2d9a7a7687
From pixman (Jeff Muizelaar)
...
Fix up the fast-path compositing operators; those are useful for
sources without alpha, but can't be used for sources with
alpha. Also, replaced fbCompositeSrcSrc_nxn with call to fbBlt as
this function must handle 1, 4, 8, 16, 24, 32 bpp objects. Would
be nice to optimize fbBlt for common cases involving 8, 16, 24 and
32bpp.
From Keith Packard.
2007-04-24 14:46:59 -04:00
Soren Sandmann Pedersen
fde4a5adf0
From xserver via pixman (Jeff Muizelaar)
...
Add some optimizations from jaymz. Also adds some compile
warnings that will hopefully go awa y as we continue merging.
2007-04-24 13:30:43 -04:00
Soren Sandmann Pedersen
13e1d5ea55
Fix format vs formatCode in previous commit
2007-04-24 12:59:18 -04:00
Soren Sandmann Pedersen
077a5d4555
Add functions fbCompositeSrcSrc_nxn() and fbCompositeTrans_0565xnx0565
...
from xserver via pixman. Add READ/WRITE and fbFinishAccess as
appropriate.
2007-04-24 12:57:55 -04:00
Peter Hutterer
09436fb7c3
Disable devices before removing, remove unrecoverable devices.
2007-04-24 23:11:44 +09:30
Peter Hutterer
c7e2ba0c9b
Use DevicePresence events to tell the client about enabled/disabled devices.
...
Include the device id in the event sent to the client.
2007-04-24 23:11:44 +09:30
Brian
ce099a9b78
fix bug in which maxKeysPerModifier wasn't getting set
2007-04-23 12:34:01 -06:00
Soren Sandmann Pedersen
b5e1f7869b
Remove #if 0'ed merge leftovers
2007-04-23 14:16:30 -04:00
Soren Sandmann Pedersen
84838268b3
Gradient fixes
...
* Port fix for bug 7685 from pixman. Patch by Carl Worth
* Add projective version of radial gradient code.
* Make sure that all Pict*Gradient types have PictGradient as prefix,
since code in various places relies on that.
2007-04-23 13:19:54 -04:00
Aaron Plattner
38d14e8589
Adjust the screen pixmap's dimensions in xf86RandR12ScreenSetSize.
2007-04-22 18:04:27 -07:00
Aaron Plattner
ca784df84e
Fix unbalanced fbGetDrawable added in commit 0a9239ec
.
2007-04-22 16:26:01 -07:00
Soren Sandmann Pedersen
55bd8668e7
Remove #if 0'ed leftovers from merge
2007-04-21 17:10:54 -04:00
Soren Sandmann Pedersen
41dd7ab067
Fix gradient walker to not reset needlessly
...
Previously the gradient walker was doing excessive resets, (such
as on every pixel in constant-colored regions or outside the
gradient with CAIRO_EXTEND_NONE). Don't do that.
Carl Worth, from pixman
2007-04-21 17:10:54 -04:00
Soren Sandmann Pedersen
c1b73f0f2a
Fixing gradient repeat mode computations in previous patch. From David
...
Turner.
2007-04-21 17:10:54 -04:00
Soren Sandmann Pedersen
38f718799c
Remove a few memory references in fbFetchTransformed
2007-04-21 17:10:54 -04:00
Soren Sandmann Pedersen
9c4b14d4f6
Integrate David Turner's gradient optimizations from pixman
2007-04-21 17:10:54 -04:00