Commit Graph

479 Commits

Author SHA1 Message Date
Maarten Maathuis
ec32878359 exa: check if the pixmap is pinned for unsupported AUX indices. 2009-03-18 19:36:18 +01:00
Maarten Maathuis
89d7b88f32 exa: allow exaModifyPixmapHeader to set sys_ptr for EXA_HANDLES_PIXMAPS
- exaModifyPixmapHeader would save sys_ptr if needed, but it would be NULL'ed afterwards.
- This is needed to support pixmaps that are not offscreen.
2009-03-13 00:44:48 +01:00
Maarten Maathuis
ce6e1771be exa: fix a serious issue in exaChangeWindowAttributes (and some more related things)
- fbChangeWindowAttributes can create pixmaps (and access them) without use preparing access.
- Also handle the destroyed pixmaps by finishing them first.
- Switch to DEST indices again in exaCreatePixmapWithPrepare, because they are obviously being rendered to.
- Also avoid calling FinishAccess on pixmaps that are destroyed (and their memory potentially invalid).
2009-03-04 16:51:10 +01:00
Maarten Maathuis
7fb68e8b31 exa: remove a few pExaPixmap checks.
- And make some fatal for a debug build.
2009-03-04 16:51:10 +01:00
Maarten Maathuis
bd2f35ef0d exa: fix unwrapping of ModifyPixmapHeader upon CloseScreen.
- Cleanup wrapping too.
2009-03-04 16:51:10 +01:00
Maarten Maathuis
da8ea41a54 exa: increase/rework safety checks in Prepare/FinishAccess. 2009-03-04 16:51:09 +01:00
Maarten Maathuis
3ea3d505e8 exa: avoid a potential Prepare/FinishAccess inbalance. 2009-03-04 16:51:09 +01:00
Maarten Maathuis
10334cf7e6 exa: simplify exaPixmapIsOffscreen
- This includes properly handling the framebuffer.
2009-03-04 16:51:09 +01:00
Maarten Maathuis
7bfad37f77 exa: fixup aux indices and ensure that the indices are used as they should be.
- In a previous patch i forgot to add a FALSE somewhere it seems.
- Rename AUX indices so the driver (think of driver managed pixmaps) can do optimisations based upon them.
- Fix one abuse of DEST index now that we have the AUX indices (same reason as above).
2009-02-27 22:26:52 +01:00
Maarten Maathuis
f028b14876 exa: whitespace 2009-02-27 17:11:47 +01:00
Maarten Maathuis
ed00515ec5 exa: fixup exaAssertNotDirty.
- Do the right thing based on prepare access.
2009-02-27 17:11:35 +01:00
Maarten Maathuis
2e88b6004f exa: minor glyphs cleanup.
- This should fix subtle coordinate bugs and make the code a bit cleaner to read.
2009-02-27 16:58:30 +01:00
Michel Dänzer
265d20068a EXA: Fix check for whether the glyph we're evicting from the cache is in use.
Since commit f07f18231a ('EXA: Allow using
exaCompositeRects also when we can't use a mask in exaGlyphs.') we were
checking the wrong set of coordinates in the buffer where glyphs to be rendered
are accumulated when no mask is used in exaGlyphs.

This fixes occasional glyph corruption which can be corrected with redraws, in
particular with Qt4.

Thanks to Maarten Maathuis for asking the right question: 'where do we protect
against evicting glyphs that are still needed?'

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-27 16:41:39 +01:00
Michel Dänzer
9d5141f7bc EXA: No longer use the driver UploadToScratch hook.
See http://bugs.freedesktop.org/show_bug.cgi?id=20213 .

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-27 16:37:28 +01:00
Michel Dänzer
4cfb36f6ad EXA: Handle separate alpha maps properly in Composite fallback, take two.
Preserve the EXA ABI by introducing a new driver flag EXA_SUPPORTS_PREPARE_AUX.
If the driver doesn't set this flag, we have to assume any Prepare/FinishAccess
driver hooks can't handle the EXA_PREPARE_AUX* indices, so we move out such
pixmaps at PrepareAccess time.

Fixes https://bugs.freedesktop.org/show_bug.cgi?id=18710 .

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-27 16:37:28 +01:00
Michel Dänzer
de022f8e63 Revert "EXA: Handle separate alpha maps properly in Composite fallback."
This reverts commit 170cf1270d.

Conflicts:

	exa/exa_render.c
2009-02-27 16:37:27 +01:00
Michel Dänzer
f07f18231a EXA: Allow using exaCompositeRects also when we can't use a mask in exaGlyphs.
This should give the full benefits of the glyph cache even when we can't use a
mask.

This also means we no longer need to scan the glyphs to see if they overlap,
we can just use a mask or not as the client asks.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-26 10:35:44 +01:00
Michel Dänzer
b26c136ee9 EXA: Stop tracking damage for pixmaps subject to ModifyPixmapHeader.
Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-26 10:35:44 +01:00
Michel Dänzer
170cf1270d EXA: Handle separate alpha maps properly in Composite fallback.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=18710 .

As this can't work without new EXA_PREPARE_AUX* indices, this requires a major
version bump, so we can also drop the UploadToScratch driver hook and
ExaOffscreenSwap*(). So this also fixes
http://bugs.freedesktop.org/show_bug.cgi?id=20213 .

Moreover, introduce EXA_DRIVER_KNOWN_MAJOR to break compilation of drivers
which may not be able to handle EXA_PREPARE_AUX*, giving instructions how to
make them build again in the #error message.

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-24 09:22:09 +01:00
Maarten Maathuis
b2ceea3635 Revert "exa: the extent of the valid region is probably much larger than that of the pending damage."
This reverts commit 97c1cbc702.

- Sorry for the thinko, pending damage is often not fragmentated.
- Should the dst region become fragmentated, you actually want to copy more to unfragmentate it.
2009-02-17 19:40:59 +01:00
Maarten Maathuis
3175646b10 exa: exaPixmapDirty should use official damage functions.
- Otherwise other users will not be notified of damage.
2009-02-17 19:25:35 +01:00
Maarten Maathuis
97c1cbc702 exa: the extent of the valid region is probably much larger than that of the pending damage.
- Since the default has been changed from nothing to everything.
2009-02-17 19:25:35 +01:00
Maarten Maathuis
caa5310e16 exa: reintroduce src rect optimisation, with a slightly higher threshold.
- I got some feedback on gtkperf line test regression.
- The increased threshold should ensure the destination optimisation is dominant.
2009-02-16 20:41:14 +01:00
Maarten Maathuis
46eeaf82e2 exa: fix performance regression from 736b6fbd2c
- The src optimisation is more aggressive and possibly harmful in light of the new initial state of pixmaps.
- There is now actually a performance improvement by almost always keeping the number of rects low.
2009-02-16 17:41:51 +01:00
Michel Dänzer
736b6fbd2c EXA: Try to prevent the valid regions from growing too many rects.
This helps contain region management overhead in some pathological cases, see
e.g. http://bugs.freedesktop.org/show_bug.cgi?id=16647 .

Signed-off-by: Michel Dänzer <daenzer@vmware.com>
2009-02-15 17:44:51 +01:00
Maarten Maathuis
1cb5afdf1e dix: always NULL pGC->tile.pixmap in Create{Scratch}GC. 2009-02-08 18:59:17 +01:00
Maarten Maathuis
6fd8fe9cb9 exa: fix exaValidateGC.
- use DEST in the createPixmap wrapper, because stipple already takes MASK (in case someone uses swappers).
- Anticipate some of the less common situations when fbValidateDrawable will access tile related pixmaps.
2009-02-08 18:43:15 +01:00
Maarten Maathuis
30daba6903 exa: All fallbacks should have a GC, remove some code.
- I did some testing with full fallbacks forced by the driver.
- I ran rendercheck, expedite and the (full) x11perf test suite.
- Thanks to ajax for pointing out this should be unneeded.
2009-02-08 02:08:13 +01:00
Maarten Maathuis
ffa72793e2 exa: Calling exaMarkSync after UTS is the drivers responsibility. 2009-02-07 21:58:17 +01:00
Maarten Maathuis
68665d78e7 exa: don't use fbCopyNtoN 2009-02-06 18:42:39 +01:00
Maarten Maathuis
2e76958d30 fb: move some code to mi 2009-02-06 18:42:39 +01:00
Maarten Maathuis
f06bde3d4b exa: create ExaCheckGetImage 2009-02-06 18:42:39 +01:00
Maarten Maathuis
258fc4b106 exa: wrap the remainder of exa_unaccel.c 2009-02-06 18:42:38 +01:00
Maarten Maathuis
6fabf24949 exa: use proper wrapping in exa.c 2009-02-06 18:42:38 +01:00
Maarten Maathuis
015c99a4ad exa: properly wrap GC functions 2009-02-06 18:42:38 +01:00
Maarten Maathuis
5e6a06fe69 exa: add GC private 2009-02-06 18:42:38 +01:00
Maarten Maathuis
5cc67ae94c exa: kill of exaImageGlyphBlt
- It serves no obvious purpose, yet it directly accesses many fb
symbols.
2009-02-06 18:42:38 +01:00
Maarten Maathuis
6483834200 exa: Remove one of the many calls directly into the fb layer. 2009-02-06 18:42:38 +01:00
Michel Dänzer
f112d6bf59 EXA: Try harder to keep current pixmap copy up to date in exaMigrateToward*.
This is probably required, but apparently not sufficient, for making migration
heuristics other than "always" work correctly again. Not that I really care
about them...
2009-02-06 18:30:43 +01:00
Michel Dänzer
3948b52389 EXA: Guard empty pending region warning by DEBUG_MIGRATE.
It isn't very useful yet while the damage layer calls us for empty operations,
mostly confuses users.
2009-02-06 11:37:53 +01:00
Michel Dänzer
639f289dcd EXA: Declare glyph cache picture as component-alpha when necessary.
Without this, rendering component-alpha glyphs may break without a mask.

Fixes http://bugs.freedesktop.org/show_bug.cgi?id=19233 .
2009-01-13 13:05:32 +01:00
Maarten Maathuis
3534a5e5d9 exa: Allow drivers to set non-NULL devPrivate.ptr for !offscreen pixmaps. 2008-12-26 16:38:27 +01:00
Maarten Maathuis
027b440d4f exa: preparing as source and finishing access as mask is a bad idea 2008-12-26 15:51:55 +01:00
Maarten Maathuis
1567b7243f exa: A more correct fix. 2008-12-21 00:08:56 +01:00
Maarten Maathuis
2db7b66863 exa: a few cleanups
- Some warnings silenced.
- Some whitespace cleanup.
2008-12-19 23:12:37 +01:00
Peter Hutterer
cb95642dc8 Remove #define NEED_EVENTS and NEED_REPLIES
A grep on xorg/* revealed there's no consumer of this define.

Quote Alan Coopersmith:
"The consumer was in past versions of the headers now located
in proto/x11proto - for instance, in X11R6.0's xc/include/Xproto.h,
all the event definitions were only available if NEED_EVENTS were
defined, and all the reply definitions required NEED_REPLIES.

Looks like Xproto.h dropped them by X11R6.3, which didn't have
the #ifdef's anymore, so these are truly ancient now."

Signed-off-by: Peter Hutterer <peter.hutterer@redhat.com>
Signed-off-by: Adam Jackson <ajax@redhat.com>
2008-12-12 11:43:32 +10:00
Paulo Cesar Pereira de Andrade
49f77fff14 Rework symbol visibility for easier maintenance
Save in a few special cases, _X_EXPORT should not be used in C source
files. Instead, it should be used in headers, and the proper C source
include that header. Some special cases are symbols that need to be
shared between modules, but not expected to be used by external drivers,
and symbols that are accessible via LoaderSymbol/dlopen.

  This patch also adds conditionally some new sdk header files, depending
on extensions enabled. These files were added to match pattern for
other extensions/modules, that is, have the headers "deciding" symbol
visibility in the sdk. These headers are:
o Xext/panoramiXsrv.h, Xext/panoramiX.h
o fbpict.h (unconditionally)
o vidmodeproc.h
o mioverlay.h (unconditionally, used only by xaa)
o xfixes.h (unconditionally, symbols required by dri2)

  LoaderSymbol and similar functions now don't have different prototypes,
in loaderProcs.h and xf86Module.h, so that both headers can be included,
without the need of defining IN_LOADER.

  xf86NewInputDevice() device prototype readded to xf86Xinput.h, but
not exported (and with a comment about it).
2008-12-03 05:43:34 -02:00
Paulo Cesar Pereira de Andrade
d6cbd4511e Export symbols defined in the sdk.
This is the biggest "visibility" patch. Instead of doing a "export"
symbol on demand, export everything in the sdk, so that if some module
fails due to an unresolved symbol, it is because it is using a symbol
not in the sdk.

  Most exported symbols shouldn't really be made visible, neither
advertised in the sdk, as they are only used by a single shared object.

  Symbols in the sdk (or referenced in sdk macros), but not defined
anywhere include:
XkbBuildCoreState()
XkbInitialMap
XkbXIUnsupported
XkbCheckActionVMods()
XkbSendCompatNotify()
XkbDDXFakePointerButton()
XkbDDXApplyConfig()
_XkbStrCaseCmp()
_XkbErrMessages[]
_XkbErrCode
_XkbErrLocation
_XkbErrData
XkbAccessXDetailText()
XkbNKNDetailMaskText()
XkbLookupGroupAndLevel()
XkbInitAtoms()
XkbGetOrderedDrawables()
XkbFreeOrderedDrawables()
XkbConvertXkbComponents()
XkbWriteXKBSemantics()
XkbWriteXKBLayout()
XkbWriteXKBKeymap()
XkbWriteXKBFile()
XkbWriteCFile()
XkbWriteXKMFile()
XkbWriteToServer()
XkbMergeFile()
XkmFindTOCEntry()
XkmReadFileSection()
XkmReadFileSectionName()
InitExtInput()
xf86CheckButton()
xf86SwitchCoreDevice()
RamDacSetGamma()
RamDacRestoreDACValues()
xf86Bpp
xf86ConfigPix24
xf86MouseCflags[]
xf86SupportedMouseTypes[]
xf86NumMouseTypes
xf86ChangeBusIndex()
xf86EntityEnter()
xf86EntityLeave()
xf86WrapperInit()
xf86RingBell()
xf86findOptionBoolean()
xf86debugListOptions()
LoadSubModuleLocal()
LoaderSymbolLocal()
getInt10Rec()
xf86CurrentScreen
xf86ReallocatePciResources()
xf86NewSerialNumber()
xf86RandRSetInitialMode()
fbCompositeSolidMask_nx1xn
fbCompositeSolidMask_nx8888x0565C
fbCompositeSolidMask_nx8888x8888C
fbCompositeSolidMask_nx8x0565
fbCompositeSolidMask_nx8x0888
fbCompositeSolidMask_nx8x8888
fbCompositeSrc_0565x0565
fbCompositeSrc_8888x0565
fbCompositeSrc_8888x0888
fbCompositeSrc_8888x8888
fbCompositeSrcAdd_1000x1000
fbCompositeSrcAdd_8000x8000
fbCompositeSrcAdd_8888x8888
fbGeneration
fbIn
fbOver
fbOver24
fbOverlayGeneration
fbRasterizeEdges
fbRestoreAreas
fbSaveAreas
composeFunctions
VBEBuildVbeModeList()
VBECalcVbeModeIndex()
TIramdac3030CalculateMNPForClock()
shadowBufPtr
shadowFindBuf()
miRRGetScreenInfo()
RRSetScreenConfig()
RRModePruneUnused()
PixmanImageFromPicture()
extern int miPointerGetMotionEvents()
miClipPicture()
miRasterizeTriangle()
fbPush1toN()
fbInitializeBackingStore()
ddxBeforeReset()
SetupSprite()
InitSprite()
DGADeliverEvent()

  SPECIAL CASES
o defined as _X_INTERNAL
	xf86NewInputDevice()
o defined as static
	fbGCPrivateKey
	fbOverlayScreenPrivateKey
	fbScreenPrivateKey
	fbWinPrivateKey
o defined in libXfont.so, but declared in xorg/dixfont.h
	GetGlyphs()
	QueryGlyphExtents()
	QueryTextExtents()
	ParseGlyphCachingMode()
	InitGlyphCaching()
	SetGlyphCachingMode()
2008-11-29 23:56:06 -02:00
Paulo Cesar Pereira de Andrade
31285d063e Make visible symbols required by xorg modules.
This patch exports all symbols required by the compilable
(in a x86 linux computer) xorg/driver/* modules.
  Still missing symbols worth mentioning are:

sunleo
	miFindMaxBand no longer available

intel	(uxa/uxa-accel.c)
	fbShmPutImage no longer available (and should have been static)

mga
	MGAGetClientPointer (should come from matrox's libhal)

  This is not a definitive "visibility" patch, as all it does is to
export missing symbols, but the modules that current don't compile,
may require more symbols once fixed, and third party drivers should
also require more symbols exported.
  A "definitive" patch should export symbols defined in the sdk.
2008-11-28 01:55:11 -02:00
Dave Airlie
08cd361234 exa: avoid doing prepare/done without intervening copies in exaFillRegionTiled
This does a precursor check to make sure the copies are required before
entering the prepare/done code.
2008-11-17 19:15:20 +10:00
Dave Airlie
8f8a9c19ad EXA: avoid copy operations if no boxes in use
Simple fix for now, I'm sure damage shouldn't be calling us with nbox = 0.
2008-11-17 10:24:39 +10:00
Michel Dänzer
21c116219c Fix typos which caused exaCompositeRects() to use an incorrect damage region.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=18065 .

Also, it can bail if either width or height of the bounding box is <= 0, not
only if both of them are.
2008-11-08 01:25:14 +01:00
Keith Packard
8767fc8d47 Wrap AddTraps in exa and damage.
This fine (and unused) function wasn't ever wrapped which made it not work
under exa.

(cherry picked from commit 06e7e1d048)
2008-10-23 09:56:45 -04:00
Adam Jackson
3891dd8924 exa: Add exaDrawableIsOffscreen() to the driver API. 2008-10-21 13:11:16 -04:00
Michel Dänzer
2188582e5e EXA: Avoid excessive syncing in PutImage 2008-10-20 09:55:24 -04:00
Maarten Maathuis
87a016ae00 exa: restore {x,y}{Src,Dst} to their original values when !done 2008-10-15 00:03:16 +02:00
Dave Airlie
717c7492a0 exa: don't call composite routines with no buffer.
We can get a case with gnome-terminal + links, where we get two arrays
of glyphs all with 0 width and 0 heights in them. If this happens
we manage to get to this case without any buffer setup and segfault.
2008-10-07 02:38:44 +10:00
Maarten Maathuis
a29af64658 exa: remove "direct" case for exa{Trapezoids,Triangles}
- By adding a small hack to the xserver i was able to easily test the performance of the normally rare direct case (using cairo).
- It turned out to be 70% slower for me (large test on an otherwise idle computer), which seems enough of a reason to remove it.
- AddTraps could also use a 2nd look, but since noone is using that it's a bit hard and less useful to test.
2008-10-03 23:59:06 +02:00
Maarten Maathuis
416870d8c9 exa: remove some excessive whitespace 2008-10-03 23:59:06 +02:00
Dave Airlie
5bf3f0fd4e exa: make sure pixmap devPrivate.ptr is NULL at create time 2008-09-26 08:59:30 +10:00
Julien Cristau
b4762c0245 exa: disable shared pixmaps
They got re-enabled in ee7c684f21
("Reimplement ShmPutImage.")
2008-09-16 17:13:42 +02:00
Dave Airlie
0466b8811c exa: drop cw.h inclusion
this doesn't seem to be needed, at least I can't see any warning without it.

I'd like to build EXA into a driver and cw.h isn't exported
2008-09-01 09:57:25 +10:00
Maarten Maathuis
f4e9a1a98f damage: choose less ambiguous function names 2008-08-31 17:46:26 +02:00
Maarten Maathuis
1861250cd7 {damage,exa}: sanitise damage
- Redo damage naming for more consistency.
- Call post submission functions only where appropriate.
- EXA can now live without it's odd damage workarounds.
2008-08-29 22:15:23 +02:00
Tomas Carnecky
ebea78cdba Prepare for array-index based devPrivates.
TODO: static indices can be made just an int; some indices
can be combined.
2008-08-28 18:05:40 -04:00
Adam Jackson
ee7c684f21 Reimplement ShmPutImage.
There's no reason to not just dispatch this straight into the GC.  As a
bonus, if you do so, damage wraps correctly, and thus swcursor works.
The side effect is it's no longer possible to override ShmPutImage with
ShmRegisterFuncs().

Also remove the (broken) damage tracking for same from EXA, since it didn't
work right, and is now superfluous.
2008-08-28 10:12:28 -04:00
Maarten Maathuis
dd1e54d6ee exa: fix thinko from 988725f32e
- the drawable of the pixmap is not the same as the original drawable (possibly a DRAWABLE_WINDOW)
2008-08-26 20:02:55 +02:00
Maarten Maathuis
7c14fdbacf exa: some minor cleanup
- Fix compile warning
- Order exa.h by source file that exports the function.
- Move the function i created earlier to private headers.
2008-08-26 17:56:40 +02:00
Maarten Maathuis
de79edbd9f exa: report damage manually for exa{Trapezoids,Triangles} when needed
- Plus a micro cleanup of unused variables.
2008-08-26 17:56:40 +02:00
Maarten Maathuis
988725f32e exa: move destination damage for internal calls to a special function
- This should improve clarity for someone who isn't familiar with the code.
2008-08-26 17:56:40 +02:00
Maarten Maathuis
ce19347680 exa_glyphs: remove useless offset 2008-08-26 17:56:39 +02:00
Maarten Maathuis
bc4cef1d0a exa: remove useless cache{X,Y}off from UploadToScreen() arguments. 2008-08-19 23:24:33 +02:00
Michel Dänzer
f227fbf74f EXA: Inline Prepare/FinishAccessWindow into only caller, ChangeWindowAttributes.
Also check the requested mask in addition to the GC state before doing work.
2008-08-19 11:36:12 +02:00
Michel Dänzer
825b3fe11d EXA: Don't use exaGlyphs if the driver doesn't provide a PrepareComposite hook.
It's buggy without Composite acceleration (leading to cropped glyphs) and not
really useful in that case anyway. The bug probably still needs to be found and
fixed for drivers that provide a PrepareComposite hook but can't accelerate
text rendering though.
2008-08-19 11:22:40 +02:00
Maarten Maathuis
421b7e8f12 exa: fix assert logic thinko from 361a9eb953
- I guess failing PrepareAccess is rare, since this a 3 year old bug.
2008-08-17 19:57:02 +02:00
Michel Dänzer
4212599c92 EXA: Make sure damage tracking code is inactive if the driver manages pixmaps.
It was always supposed to be like that... It was only recently pointed out (in
a rather convoluted way) that it was not in fact the case.
2008-08-08 12:17:58 +02:00
Daniel Stone
e882ee7056 EXA: Remove unnecessary #includes
There wasn't actually any font code here, so no problem.
2008-08-05 01:17:32 +03:00
Michel Dänzer
148175fb8b EXA: Do still return FALSE if the driver PrepareCopy hook failed...
Thanks to Stuart Bennett for pointing out the problem on IRC.
2008-07-31 16:54:33 +02:00
Michel Dänzer
64ebeeb526 EXA: Fix exponential growth logic for GXcopy tiled fills.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=16908 .
2008-07-31 10:58:52 +02:00
Michel Dänzer
8405c25d9d EXA: Simplify exaFillRegionTiled() control flow.
Also only call REGION_TRANSLATE() when necessary.
2008-07-31 10:55:44 +02:00
Michel Dänzer
a3afa6f2fb EXA: Optimize GXcopy tiled fills. 2008-07-30 18:30:37 +02:00
Michel Dänzer
37087bc106 EXA: Replace open coded CopyArea logic with GC op call. 2008-07-30 18:27:33 +02:00
Alan Hourihane
528b4e36ad Set driverPriv immediately on CreatePixmap.
If it's NULL anyway, we bail, if not, it lets
ModifyPixmapHeader know about the private.
2008-06-18 22:34:02 +01:00
Michel Dänzer
6b96281100 EXA: Fix exaGetPixmapFirstPixel() crash if the driver has a CreatePixmap hook.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=16243
2008-06-06 11:01:03 +02:00
Michel Dänzer
29586101dc EXA: Only record damage generated by rendering operations.
Recording damage from other operations (e.g. creating a client damage record)
may confuse the migration code resulting in corruption.

Option "EXAOptimizeMigration" appears safe now, so enable it by default. Also
remove it from the manpage, as it should only be necessary on request in the
course of bug report diagnostics anymore.
2008-05-24 20:01:41 +02:00
Michel Dänzer
f6d61d3d86 EXA: Don't migrate twice in exaImageGlyphBlt.
exaPrepareAccess already handles migration.
2008-05-24 20:01:36 +02:00
Michel Dänzer
d733043982 EXA: Don't suggest exaCopyDirty be inlined.
Leave the decision to the compiler toolchain.
2008-05-24 20:01:31 +02:00
Michel Dänzer
3baf3b42e0 EXA: Specify region used for source pixmap migration in exaCopyNtoN.
Avoids excessive migration overhead in some pathological cases. See
http://bugs.freedesktop.org/show_bug.cgi?id=15845 .
2008-05-24 20:01:21 +02:00
Michel Dänzer
3b8d53452c EXA: Fall back in CompositeRects if the driver can't accelerate Composite.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=15780 .
2008-05-01 15:10:52 +02:00
Michel Dänzer
8349732a67 EXA: Try to accelerate non-antialiased text via the glyph cache as well.
Treat 1 bit glyphs and masks as PICT_a8 in the glyph cache. We're not able to
accelerate them otherwise.
2008-04-28 21:09:35 +02:00
Michel Dänzer
a65d530040 EXA: Accumulate glyphs whenever possible, for full benefits of the glyph cache. 2008-04-28 21:03:12 +02:00
Michel Dänzer
e7eaac59c4 EXA: Glyph cache upload tweaks.
Track damage after using UploadToScreen directly.

Don't waste any effort on empty glyphs.
2008-04-28 21:00:55 +02:00
Owen Taylor
cc08c06665 EXA: Use UploadToScreen() for uploads to glyph cache
When possible, use UploadToScreen() rather than CompositePicture()
to upload glyphs onto the glyph cache pixmap. This avoids allocating
offscreen memory for each glyph making management of offscreen
areas much more efficient.
2008-04-28 21:00:55 +02:00
Owen Taylor
13fd225630 EXA: Clean up debug messages 2008-04-28 21:00:55 +02:00
Owen Taylor
fcb5949928 EXA: Fix overlapping glyphs in glyph cache
Allocate each cache at a different vertical position in the
per-format pixmap. Fix width/height confusion when choosing
the cache for a glyph.
2008-04-28 21:00:54 +02:00
Owen Taylor
40eb14c948 EXA: Add exaCompositeRects()
Add a function to composite multiple independent rectangles
from the same source to the same destination in a single
operation: this is useful for building a glyph mask.
2008-04-28 21:00:54 +02:00
Owen Taylor
54184110f6 EXA: Use a single large glyph cache pixmap
Add back exaGlyphs(); the new version copies the glyph images
onto a single large glyph pixmap and draws from their to the
destination surface. This reduces the management of small
offscreen areas and will allow us to avoid texture unit setup
between each glyph.
2008-04-28 21:00:54 +02:00
Tilman Sauerbeck
3f081b4de5 EXA: Set pixmap->accel_blocked on the screen pixmap, too. 2008-04-21 23:07:39 +02:00
Tilman Sauerbeck
26c1801a27 EXA: Update sys_pitch/fb_pitch in exaModifyPixmapHeader.
exaModifyPixmapHeader now also only evaluates arguments that have a
meaningful value.
2008-04-21 23:07:39 +02:00
Michel Dänzer
6c95fae1e9 EXA: Offscreen memory eviction improvements.
* Make sure available areas are considered to have no eviction cost. This seems
  to help for https://bugs.freedesktop.org/show_bug.cgi?id=15513 but I'm afraid
  that may just be coincidence.
* Only calculate eviction cost of each area once for each eviction pass.
  Safeguard against potential (though unlikely) division by zero.
* Cosmetic enhancements: Name eviction cost related variables 'cost' instead of
  'score' to emphasize that smaller values are better, update Doxygen file
  comment to the way eviction works now.
2008-04-21 10:45:11 +02:00
Tilman Sauerbeck
886af8f384 EXA: Avoid some fallbacks in exaCopyNtoN.
In some cases we can still do the copying in hardware even if the
dimensions of the pixmaps are out of range. This is true when the boxes
that we're to copy are all in the card's range.
2008-04-17 14:28:31 +02:00
Tilman Sauerbeck
f133d85778 EXA: Update pixmaps' accel_blocked field in ModifyPixmapHeader. 2008-04-14 12:48:13 +02:00
Tilman Sauerbeck
bb8868540f EXA: Teach exaCompositeFallbackPictDesc() about x8r8g8b8. 2008-04-14 12:48:13 +02:00
Owen W. Taylor
6b1a27023e EXA: Fix TS origin computation when implementing RenderComposite with tiling. 2008-04-07 11:01:09 +02:00
Fredrik Höglund
8074676d2d EXA: Optimize the eviction scanning loop in exaOffscreenAlloc.
Reduce the cost of the inner loop, by keeping a set of pointers to the
first and the last areas in the series, subtracting the cost of the first
area from the score, and adding the cost of the last area while walking
the list.  This commit also moves the scanning loop from exaOffscreenAlloc
into a separate function.

Idea by Michel Dänzer.
2008-03-31 21:24:59 +02:00
Fredrik Höglund
93d876891d EXA: Improve the algorithm used for tracking offscreen pixmap use.
Replace the current score keeping algorithm with a rolling counter that's
incremented in ExaOffscreenMarkUsed, with the previous value being stored
in the area.  exaOffscreenAlloc uses the difference between the counter
value and the value in the area when deciding which area to evict.
It now also takes the size of the areas into account, and favors evicting
smaller areas.

The credit for these ideas goes to Michel Dänzer.
2008-03-31 21:15:50 +02:00
Maarten Maathuis
f6d51797a5 EXA: Let the driver decide what repeat/extend types are supported. 2008-02-27 23:44:39 +01:00
Eric Anholt
5cb9e15562 EXA: Fix Render acceleration in copy and tiling cases.
Code shuffling in a634c9b034 broke this by
leaving pSrcPixmap = NULL.
2008-02-20 10:36:06 -08:00
Bernardo Innocenti
e6ea3147bf exa: make the prototype for exaGetPixmapFirstPixel() public
This fixes a warning in amd_drv which is using it.

Signed-off-by: Bernardo Innocenti <bernie@codewiz.org>
2008-01-14 13:09:13 -05:00
Alan Hourihane
57468a696e Fix Line drawing with CapNotLast set in PolySegment. 2008-01-04 12:37:55 +00:00
Eamon Walsh
2d17f47cc7 Merge branch 'master' into XACE-SELINUX
Conflicts:

	hw/xnest/Pixmap.c
	include/dix.h
2007-11-19 18:10:46 -05:00
Ben Skeggs
c89b543198 exa: set driverPriv to NULL before it might get used later with garbage 2007-11-17 18:20:49 +10:00
Eamon Walsh
1603130236 Merge branch 'master' into XACE-SELINUX
Conflicts:

	Xext/xace.c
	Xext/xace.h
2007-11-14 13:35:50 -05:00
Michel Dänzer
9bee1c6912 EXA: Disable problematic optimization of dest pixmap migration by default.
Also add some code comments about these optimizations.
2007-11-07 19:01:24 +01:00
Eamon Walsh
a52c9b2a59 Merge branch 'master' into XACE-SELINUX
Conflicts:

	dix/dispatch.c
	dix/property.c
	hw/xfree86/common/xf86VidMode.c
	include/xkbsrv.h
	render/glyph.c
	xkb/xkbActions.c
2007-11-05 19:08:36 -05:00
Daniel Stone
1179ddea94 EXA: Remove usage of alloca
Replace with heap allocations.
2007-11-05 14:36:36 +00:00
Daniel Stone
3b77689266 EXA: Remove usage of alloca
Replace with heap allocations.
2007-11-05 14:34:43 +00:00
Aaron Plattner
f2e310132f Add CreatePixmap allocation hints.
These hints allow an acceleration architecture to optimize allocation of certain
types of pixmaps, such as pixmaps that will serve as backing pixmaps for
redirected windows.
2007-11-04 16:11:28 -08:00
Eamon Walsh
b633d54b94 Merge branch 'master' into XACE-SELINUX
Conflicts:

	GL/glx/glxscreens.c
	hw/xnest/Screen.c
	render/glyph.c
	render/glyphstr.h
	render/render.c
2007-10-25 12:19:30 -04:00
Pierre Willenbrock
d502521c36 EXA: Fix off-by-one in polyline drawing. 2007-10-23 16:45:13 +02:00
Eric Anholt
4b14c9a9cd Replace calls to Glyphs screen hook with CompositeGlyphs and remove dead code.
Not all of the DDX/miext Glyphs hook implementations have been removed, but
they should be.
2007-10-19 16:34:54 -07:00
Eric Anholt
a3a95d3475 Merge branch 'master' into glyph-pixmaps
Conflicts:

	configure.ac
2007-10-19 15:44:17 -07:00
Michel Dänzer
ce50bfd336 EXA: Skip empty glyphs. 2007-10-18 17:44:48 +02:00
Michel Dänzer
5d74416740 EXA: Don't attempt to move in pixmaps that can't be accelerated.
Fixes https://bugs.freedesktop.org/show_bug.cgi?id=12815 .
2007-10-18 17:44:14 +02:00
Eamon Walsh
c3f7b86255 Merge branch 'master' into XACE-SELINUX
Conflicts:

	os/access.c
2007-10-17 15:00:54 -04:00
Dodji Seketeli
6aab6a6e2a EXA: dude, don't validation driver hooks twice ! 2007-10-08 13:25:38 +02:00
Michel Dänzer
3549a12823 EXA: Disable 1x1 pixmap fill optimization for drivers that handle pixmaps.
This reverts commit 1365aeff54.

It defeated the optimization for drivers that don't provide a CreatePixmap
hook. The optimization makes no sense for drivers that do anyway, so disable
it for them completely.
2007-10-04 11:44:03 +02:00
Dodji Seketeli
604ebb5a6d Merge Xephyr-XV/GL stuff into master 2007-10-03 16:14:08 +02:00
Dave Airlie
1365aeff54 exa: direct access to the pixmap sys ptr is bad if the pixmap isn't mapped 2007-10-03 12:01:42 +10:00
Dave Airlie
566dd3b7d7 exa: add new flags to denote driver handles all pixmap migration/hiding 2007-10-03 12:01:37 +10:00
Dave Airlie
1df1fee82d exa: make sure we set fb_ptr to NULL 2007-10-03 11:56:04 +10:00
Eric Anholt
439edc768e Merge branch 'glyph-pixmaps'
Conflicts:

	configure.ac
	exa/exa_render.c
2007-10-02 12:14:04 -07:00
Dodji Seketeli
95fadbd402 Add basic Exa driver entry point validation.
In exaDriverInit(), quickly check if the programmer forgot to set some
mandadory driver hooks, in that case display a meaningful error message.
2007-10-02 16:45:01 +02:00
Dave Airlie
dc90d50008 exa: increase minor version number for pixmap allocation hooks 2007-10-02 14:05:07 +10:00
Kristian Høgsberg
f15af2ae60 exa: only setup offscreen allocator if driver doesn't provide CreatePixmap 2007-10-02 14:03:48 +10:00
Dave Airlie
ffb58f4fa8 exa: add hooks for drivers to take over pixmap allocation
This adds hooks for the driver to access Create/DestroyPixmap and ModifyPixmapHe
ader.

It allocates a 0 sized pixmap using fb and calls the driver routine to do
work of allocating the actual memory.

ModifyPixmapHeader is mainly required for hooking the screen pixmap which
isn't create by normal methods
2007-10-02 14:03:39 +10:00
Dave Airlie
31555af000 exa: add a pixmap private pointer for drivers to retrieve. 2007-10-02 13:57:55 +10:00
Tilman Sauerbeck
cfe549d1ba EXA: Added back the maxPitchPixels initialization code.
This doesn't add real value yet, but it will be useful once I add code
that splits large render operations into smaller parts if necessary.
2007-09-27 22:36:52 +02:00
Michel Dänzer
6d5c1e0d89 EXA: Remove bogus pitch checks.
exaCreatePixmap should handle all cases correctly.
2007-09-27 13:35:23 +02:00
Michel Dänzer
598698678b EXA: Punt for all fallbacks in exaFillRegion*.
Now that PaintWindow is gone, all callers already handle fallbacks.
2007-09-27 13:33:52 +02:00
Michel Dänzer
006f652505 EXA: Make sure tile offsets passed to drivers are never negative.
Thanks to Björn Steinbrink for pointing out the problem on IRC.
2007-09-27 13:30:58 +02:00
Michel Dänzer
da7d9aa1fb EXA: Tile offscreen pixmap coordinate offsets are always 0. 2007-09-27 13:30:48 +02:00
Michel Dänzer
d6f4764bf5 EXA: Remove some clearly bogus code from exaCopyNtoN.
Not sure what I was thinking when I wrote this... it would cause the box
coordinates to be off for exaCopyNtoNTwoDir or fallbacks.

Thanks to Tilman Sauerbeck for pointing out the problem on IRC and testing the
fix.
2007-09-27 13:29:07 +02:00
Michel Dänzer
c7d6d1f589 EXA: Punt on fallback case not handled correctly in exaFillRegionTiled.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=12520 .
2007-09-27 13:20:45 +02:00
Eamon Walsh
27612748e0 Merge branch 'master' into XACE-SELINUX
Conflicts:

	dix/devices.c
2007-09-26 07:47:29 -04:00
Tilman Sauerbeck
56ffc381d3 EXA: Removed duplicated exaGetDrawablePixmap() calls. 2007-09-26 16:52:27 +02:00
Tilman Sauerbeck
ae8b4f7dcf EXA: Added pitch limitations.
Drivers can now specify the max pitches that the accelerator supports.
2007-09-25 16:43:06 +02:00
Tilman Sauerbeck
edd8817083 EXA: Check ABI compatibility earlier.
If the driver isn't compatible to the server, all bets are off anyway wrt
the contents of the fields that we're validating, which can lead to bogus
error messages.
2007-09-22 13:52:56 +02:00
Eamon Walsh
97c150b61b Merge branch 'master' into XACE-SELINUX
Conflicts:

	afb/afbpntwin.c
	afb/afbscrinit.c
	afb/afbwindow.c
	cfb/cfb.h
	cfb/cfballpriv.c
	cfb/cfbscrinit.c
	cfb/cfbwindow.c
	configure.ac
	fb/wfbrename.h
	hw/xfree86/xf4bpp/ppcIO.c
	hw/xfree86/xf4bpp/ppcPntWin.c
	hw/xfree86/xf4bpp/ppcWindow.c
	hw/xfree86/xf8_32bpp/cfbscrinit.c
	mfb/mfb.h
	mfb/mfbpntwin.c
	mfb/mfbscrinit.c
	mfb/mfbwindow.c
	mi/miexpose.c

Note: conflicts caused by devPrivates rework vs. paintwindow changes.
2007-09-19 07:25:55 -04:00
Michel Dänzer
547ad2125e EXA: Make sure driver hooks get correct offscreen offsets from exaCopyDirty.
This should ensure the driver UploadTo/DownloadFromScreen hooks can always
work as intended.
2007-09-17 20:40:13 +02:00
Alan Hourihane
56cc24ffb2 EXA: Don't crash in ExaCheckPolyArc.
See https://bugs.freedesktop.org/show_bug.cgi?id=12286 .
2007-09-17 20:40:06 +02:00
Michel Dänzer
7ac89060e0 EXA: Don't crash in exaGetImage when swapped out. 2007-09-17 20:39:59 +02:00
Tilman Sauerbeck
3876c6c805 EXA: Fixed compiler warnings. 2007-09-17 18:47:45 +02:00
Tilman Sauerbeck
d5738ff2e0 EXA: Added missing exaPrepare/FinishAccess calls in ExaCheckPushPixels. 2007-09-17 17:42:55 +02:00
Eric Anholt
e4d11e58ce Remove the PaintWindow optimization.
This was an attempt to avoid scratch gc creation and validation for paintwin
because that was expensive.  This is not the case in current servers, and the
danger of failure to implement it correctly (as seen in all previous
implementations) is high enough to justify removing it.  No performance
difference detected with x11perf -create -move -resize -circulate on Xvfb.
Leave the screen hooks for PaintWindow* in for now to avoid ABI change.
2007-09-13 00:08:53 +00:00
Michel Dänzer
e8093e15c7 EXA: Exclude bits that will be overwritten from migration in exaCopyNtoN.
Also plug a region leak in exaPolyFillRect.
2007-09-07 18:41:33 +02:00
Michel Dänzer
e81af8ba64 EXA: exaFillRegion{Solid,Tiled} improvements.
Use region to exclude bits that will be overwritten from migration.

Also make exaFillRegionSolid use the same logic as exaFillRegionTiled.
2007-09-07 18:41:33 +02:00
Michel Dänzer
5f7da4da8d EXA: Use exaShmPutImage for pushing glyphs to scratch pixmap in exaGlyphs. 2007-09-07 18:41:33 +02:00
Michel Dänzer
be922b3048 EXA: exa(Shm)PutImage improvements.
Improve exaShmPutImage performance and reuse its core in exaPutImage as it
seems faster than the previous code when the driver doesn't provide an
UploadToScreen hook.

Make sure all damage records are notified of the damage incurred by actual
ShmPutImage calls.

Remove superfluous manual damage tracking for actual PutImage calls.
2007-09-07 18:41:33 +02:00
Michel Dänzer
ea92ea4156 EXA: exaGetImage improvements.
Use the new migration infrastructure to cache FB bits we need in the system
copy, for the benefit of repeated calls.
2007-09-07 18:41:32 +02:00
Michel Dänzer
aa2ed73e0e EXA: Remove superfluous manual damage tracking.
These should all be covered by damage wrappers.
2007-09-07 18:41:32 +02:00
Michel Dänzer
a634c9b034 EXA: RENDER improvements.
Exclude bits that will be overwritten from migration.

Use exaGlyphs even when Composite can't be accelerated, to avoid PolyFillRect
roundtrip via offscreen memory.

Initialize mask pixmap in exaGlyphs in FB in addition to system if the driver
provides Composite hooks to avoid migration overhead.

Remove manual damage tracking where superfluous.
2007-09-07 18:41:32 +02:00
Michel Dänzer
1f457ff3db EXA: Improvements for 1x1 pixmaps.
Initialize system and FB copy in exaFillRegionSolid and adapt
exaGetPixmapFirstPixel to the new migration infrastructure.

This should mostly eliminate migration overhead for these, whether they are
used for acceleration or fallbacks.
2007-09-07 18:41:31 +02:00
Michel Dänzer
489bc7551f EXA: exaImageGlyphBlt improvements.
As we can't actually accelerate anything interesting here, just migrate out
once and call fbSolidBoxClipped instead of taking a round trip via offscreen
memory with exaSolidBoxClipped.

Reuse pending damage region for extents and to prevent any actual migration of
pixmap contents when we're overwriting the whole pending damage region.

Remove superfluous manual damage tracking.
2007-09-07 18:41:31 +02:00
Michel Dänzer
2e0895a4ba EXA: Improvements for trapezoids and triangles.
Only migrate once in exaTrapezoids/Triangles instead of every time in
exaRasterizeTrapezoid/AddTriangles. Adapt manual damage tracking to new
infrastructure.

Also move definition of NeedsComponent() closer to where it's used.
2007-09-07 18:41:31 +02:00
Michel Dänzer
6c9d7ed61b EXA: Hide pixmap pointer outside of exaPrepare/FinishAccess whenever possible.
We finally want to catch all cases where the pixmap pointer is dereferenced
outside of exaPrepare/FinishAccess.

Also fix a couple of such cases exposed by this change.
2007-09-07 18:41:30 +02:00
Michel Dänzer
962eddd7a2 EXA: Support partial migration of pixmap contents between Sys and FB.
The initiator of migration can pass in a region that defines the relevant area
of each source pixmap or the irrelevant area of the destination pixmap. By
default, the pending damage region is assumed relevant for the destination
pixmap, and everything for source pixmaps.

Thanks to Jarno Manninen for reassuring me that my own ideas for this were
feasible and for providing additional ideas.
2007-09-07 18:41:30 +02:00
Michel Dänzer
5c7ee3f47f EXA: Track valid bits in Sys and FB separately.
Also consolidate exaCopyDirtyToFb/Sys.
2007-09-07 18:41:29 +02:00
Michel Dänzer
8cfcf9973c EXA: Migrate out pixmap in exaPrepareAccess.
Also fix exaFinishAccessGC not to use the same index for tile and stipple.
2007-09-07 18:41:29 +02:00
Dave Airlie
e510a77ba4 EXA: Add a couple of missing exaPrepare/FinishAccess calls. 2007-09-07 18:41:29 +02:00
Eamon Walsh
4017d31902 devPrivates rework: since API is already broken, switch everything
over to new system.

Need to update documentation and address some remaining vestiges of
old system such as CursorRec structure, fb "offman" structure, and
FontRec privates.
2007-08-28 09:28:25 -04:00
Adam Jackson
ae7f71a8b3 Implement core protocol backing store exclusively in terms of Composite.
Composite's automatic redirection is a more general mechanism than the
ad-hoc BS machinery, so it's much prettier to implement the one in terms
of the other.  Composite now wraps ChangeWindowAttributes and activates
automatic redirection for windows with backing store requested.  The old
backing store infrastructure is completely gutted: ABI-visible structures
retain the function pointers, but they never get called, and all the
open-coded conditionals throughout the DIX layer to implement BS are gone.

Note that this is still not a strictly complete implementation of backing
store, since Composite will throw the bits away on unmap and therefore
WhenMapped and Always hints are equivalent.
2007-08-25 15:08:20 -04:00
Michel Dänzer
b9a806f0b3 exaPolyFillRect: pGC->alu doesn't matter with a single rectangle. 2007-08-24 19:04:55 +02:00
Michel Dänzer
d0f0d1092c exaGetImage: Don't migrate pixmap out of FB with no DownloadFromScreen hook.
Based on the assumption that GetImage is relatively rare, so the overhead of
the migration is probably bigger than any potential savings.
2007-08-24 14:10:13 +02:00
Michel Dänzer
5d9e2c2821 EXA: Improve ShmPutImage.
Share as much code with exaPutImage as possible, and fall back to fbShmPutImage
when that fails.
2007-08-24 14:03:14 +02:00
Michel Dänzer
c19f227b46 EXA: Only mark offscreen memory as used when it really is. 2007-08-24 13:05:52 +02:00
Eric Anholt
bc2d516f16 Fix overly-restrictive integer overflow check in EXA pixmap creation.
The result was that at 32bpp, pixmaps of width 8192 or greater couldn't be
created, due to treating a pitch value as a width.
2007-08-22 08:59:10 -07:00
Fredrik Höglund
3c448b0eb6 EXA: Fix a couple of logic errors in exaGetPixmapFirstPixel.
The fb pointer would be left uninitialized when exaPixmapIsOffscreen
returned false. When it returned true and the pixmap was damaged,
fb would be initialized from the pixmap's devPrivate.ptr before the
exaDoMigration and exaPrepareAccess calls, at which point
devPrivate.ptr would still be pointing at offscreen memory.
2007-08-18 19:02:18 +02:00
Fredrik Höglund
daee59b170 EXA: Wrap Trapezoids to prevent excessive migration of the alpha pixmap.
miTrapezoids creates an alpha pixmap and initializes the contents
using PolyFillRect, which causes the pixmap to be moved in for
acceleration. The subsequent call to RasterizeTrapezoid won't be
accelerated by EXA, which causing the pixmap to be moved back out
again.

By wrapping Trapezoids and using ExaCheckPolyFillRect instead of
PolyFillRect to initialize the pixmap, we avoid this roundtrip.
2007-08-16 23:49:07 +02:00
Carl Worth
0a71e1542a Create a Picture as well as a Pixmap at the time of AllocateGlyph
This avoids some inefficiency in creating a temporary Picture
for every glyph at rendering time. My measurements with an i965
showed the previous patch causing a 10-15% slowdown for NoAccel
and XAA cases, (while providing an 18% speedup for EXA).

With this change, the NoAccel and XAA performance regression is
eliminated, and the overall EXA speedup, (before any of the
glyphs-as-pixmaps work), is now 32%.
2007-08-02 22:49:56 -07:00
Carl Worth
a2af34d5a8 Use per-screen Pixmaps for glyphs
Instead of system-memory data which prevents accelerated
compositing of glyphs, (at least without forcing an upload
of the glyph data before compositing).
2007-08-02 22:49:56 -07:00
Adam Jackson
486fd4145a exaGetPixmapFirstPixel: avoid framebuffer readbacks if possible.
If the pixel in framebuffer memory isn't modified since we uploaded it, we
can just read from the system memory copy, wihch avoids both a readback and
an accelerator stall.

In principle this function is still wrong, and all the framebuffer pixel
access should be going through (w)fb so we can get pixel layout corrections.
2007-07-27 13:10:39 -04:00
Dodji Seketeli
5657fb065c exaDriverInit: Fail if pScreenInfo or a member of it is invalid.
EXA may attempt to use the invalid value and crash otherwise.
2007-07-17 12:12:02 +02:00
Dodji Seketeli
2a75c77497 ExaOffscreenMarkUsed: Don't crash when there's no offscreen memory. 2007-07-03 11:00:29 +02:00
Michel Dänzer
c9b79a3558 exaPolyFillRect: Don't track damage explicitly.
All callers should already do it.

Also don't leak pReg.
2007-06-20 18:56:06 +02:00
Michel Dänzer
87966c5d28 exaGetImage: Don't migrate or try to accelerate for 1x1.
This is mainly to avoid wasting effort for XSync(), but just reading a single
pixel directly is probably faster than DownloadFromScreen anyway. Though in
light of the latter, even larger thresholds might be useful.

Also move the swappedOut check before the migration checks because migration
can't actually occur when swapped out.
2007-06-20 18:46:03 +02:00
Michel Dänzer
cc648e609d EXA: Export ExaOffscreenMarkUsed.
Can be used to inform EXA that an offscreen area is used outside of EXA.
2007-05-22 10:51:56 +02:00
Jesse Barnes
1b3a0508a7 Fix documentation for Copy hook -- it can copy memory to the scanout
buffer too.
2007-05-06 01:30:59 -07:00
Michel Dänzer
3c91a993e8 EXA: Fix OffscreenValidate build with DEBUG_OFFSCREEN enabled. 2007-04-29 23:49:41 +02:00
Michel Dänzer
a261e13250 EXA: Remove DrawableDirty.
Convert the remaining callers to PixmapDirty.
2007-04-29 23:49:35 +02:00
Michel Dänzer
b1b6674a91 EXA: FillSpans improvements.
* Don't need to track damage.
* Always migrate for fallbacks.
2007-04-29 23:49:28 +02:00
Michel Dänzer
584697a223 EXA: SolidBoxClipped improvements.
* Centralize handling of fallbacks and damage tracking.
* Always migrate for fallbacks.
2007-04-29 23:49:09 +02:00
Michel Dänzer
982d7c2c0b EXA: CopyNtoN improvements.
* Centralize handling of fallbacks and damage tracking.
* Always migrate for fallbacks.
2007-04-29 23:48:59 +02:00
Michel Dänzer
d2245386ee EXA: GetImage improvements.
Only migrate when appropriate. In particular, don't migrate to offscreen in the
no-fallback case as copying from system memory should usually be as fast if not
faster than DownloadFromScreen, in particular if the bits need to be uploaded
to offscreen first.
2007-04-29 23:48:31 +02:00
Michel Dänzer
0880aaac9c EXA: PutImage improvements.
* Migrate for fallbacks when appropriate.
* Add damage tracking in ExaCheckPutImage.
2007-04-29 23:48:19 +02:00
Michel Dänzer
7fca169011 EXA: ImageGlyphBlt improvements.
* Don't waste effort on invisible glyphs.
* Only track damage for bounding box instead of each glyph separately.
* Always migrate for fallbacks.
2007-04-29 23:48:11 +02:00
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