14aafc258c
Change EXA so that exaMoveOutPixmap() retains the framebuffer copy of the
...
pixmap, and damage is tracked so that a later exaMoveInPixmap won't
result in an upload if no upload is necessary. This will likely improve
the performance of the "Always" migration scheme significantly, and is
a step in the path to more exact damage tracking between framebuffer
and system memory.
2006-03-16 18:43:55 +00:00
d0d336efd5
Add support for ZX2 PCI-E local bus adaptors. (Alex Williamson, HP)
...
Use soft timeout register to avoid MCAs when probing for non-existent local
bus adaptors on ZX2. (Alex Williamson, HP)
2006-03-16 16:29:17 +00:00
175980580e
Bump to requiring fixesproto >= 4.0 and compositeproto >= 0.3.
2006-03-15 23:05:53 +00:00
6fe377af5a
- OpenBSD needs -Wl,-export-dynamic to export symbols from main executable
...
to modules.
- Probe for OpenBSD aperture driver and define HAS_APERTURE_DRV
accordingly.
2006-03-15 21:25:38 +00:00
21f7f2fb11
Enable correct handling of the BTS instruction (opcode 0f ab) The code was
...
there but #ifdefed out. Insead of BTS, BT was executed. This patch
enables the BTS function and hooks it up the the correct opcode. (ATI
Technologies Inc.)
2006-03-15 18:43:32 +00:00
b726aa502a
Update to build against Mesa CVS HEAD.
2006-03-15 18:37:44 +00:00
c74464d92c
Don't let pinned pixmaps get migrated in when using the "Always" migration
...
scheme. This notably keeps the visible screen from getting migrated in
to a new location in framebuffer.
Reported by: Michel Dänzer.
2006-03-15 16:59:45 +00:00
b9c43cde1e
Coverity #1042 , 1043: Nuke some dead variables.
2006-03-15 16:56:10 +00:00
5e106a71b9
Coverity #807 : Fix a memory leak in XFixesExpandRegion.
2006-03-15 16:49:04 +00:00
a3ef63696c
Coverity #490 : Fix a range check in xf86vidmode extension.
2006-03-15 16:36:31 +00:00
152090ce44
Coverity #487 : Check version number correctly.
2006-03-15 16:33:12 +00:00
72cc630725
Coverity #491 : Check version number correctly.
2006-03-15 16:32:05 +00:00
460f2ea4a5
Coverity #794 : Fix a highly unlikely memory leak.
2006-03-15 16:21:04 +00:00
116d158e85
Coverity #269 : Compare the requested ABI class against the ABI class of the
...
module, not the module class.
2006-03-15 16:16:24 +00:00
d8221a9b70
Coverity #484 : Fix an off-by-one in module refcounting.
2006-03-15 16:11:34 +00:00
6bb2dc02a7
Coverity #337 : Remove useless NULL check.
2006-03-15 16:01:47 +00:00
1e5c0842af
Coverity #1053 : Nuke a dead variable.
2006-03-15 15:34:57 +00:00
7314d16cde
Fix DRIExtensionInit() to not register callbacks when it hasn't been
...
initialized for the current server generation. Fixes a problem where it
would use stale private index and blow up in colorful ways if no driver
called DRIScreenInit() on the second generation (which happens due to a
bug in radeon that i'll fix separately). Note: clearing the index in
DRIReset() wouldn't work as DRIReset() is called before the
CloseScreen() chain
2006-03-15 03:18:42 +00:00
02d80a0de9
Make xf86 linear allocator smarter when dealing with alignment constraints
...
when falling back to X/Y allocations. Fixes various problems of Xv
allocation failures, notably with "nv" driver.
2006-03-15 03:12:32 +00:00
c1601717d5
Add a new migration scheme, "always", which will move pixmaps to their
...
desired location always (unless they don't fit in FB, in which case
they all get moved out for software rendering). The default remains as
before, but can be controlled by the MigrationHeuristic xorg.conf
option (which is intentionally not documented, as it may be
short-lived). This is part of the exa-damagetrack work, which appears
stable in testing with fakexa, unlike the work as a whole.
2006-03-15 01:20:08 +00:00
a90cff266c
Add more doxygen documentation, including notes on WaitMarker() and
...
MarkSync() that I noticed were needed while reading the VIA driver.
2006-03-15 00:13:52 +00:00
693e42114f
Move migration logic to a new function, exaDoMigration(). This is largely a
...
manual conversion to allow for different migration schemes to be
implemented reasonably, but does include some minor improvements such
as accounting for pinned pixmaps not being acceleratable, and for our
current GetImage and GetSpans not being accelerated.
2006-03-14 21:30:12 +00:00
d309054780
Pull code for getting the (0,0) pixel from a pixmap out to a separate
...
function, since it gets repeated (with bad error handling, in one
case).
2006-03-14 20:38:06 +00:00
01aa209f20
Bail out early if screen doesn't support DRI.
2006-03-14 19:32:27 +00:00
0cc34266d6
Updated ChangeLog for my latest composite and xfixes changes.
2006-03-13 22:43:42 +00:00
e5956f49b2
First batch of AIGLX fixes from David Reveman.
...
Add getter for Mesa provider.
Export this for Xgl.
Move resource tracking out of drawable constructor to allow wrapping.
Use corrent reply size #define.
Add this function. (DoGetDrawableAttributes): Fix array length.
2006-03-13 01:54:59 +00:00
eb63e50d95
Fix build when AIGLX is false.
2006-03-12 17:14:03 +00:00
9ed3463450
Improve doxygen formatting, and attempt to clarify the 1:1 ratio of
...
successful PrepareCopy()s to DoneCopy()s.
2006-03-12 03:04:52 +00:00
9a7fba5fd0
Make exaCopyNtoNTwoDir() call DoneCopy() at the end of each string of
...
consecutive Copy() calls (rather than exactly once at the end of the
function).
Reviewed by: jbarnes
2006-03-12 03:02:26 +00:00
c3342c8000
Merge accel_indirect branch to HEAD.
2006-03-12 00:11:34 +00:00
b1b731c286
Fix buffer size checks to prevent 2-byte buffer overflows. (Coverity #480 ,
...
#481 , #482 , #483 )
2006-03-11 02:43:51 +00:00
fc0772de36
Add HAS_MMAP for Xvfb
...
Fix Xvfb option parsing to exit on bad arguments, not just issue error
messages and continue on. (Coverity #492 )
2006-03-11 02:10:14 +00:00
f2ecbb3018
Pass sizeof the correct buffer to XmuSnprintf. (Coverity #489 )
2006-03-11 01:58:32 +00:00
d695579848
If fakexa is enabled, create a larger buffer in the Ximage, but keep the
...
same width/height for front-buffer drawing. The fakexa code then uses
this extra space for offscreen pixmaps. Note that this tones down the
absurdity of fakexa's offscreen pixmap alignment requirements (odd
alignment is too weird, so stick with "24", which is still strange but
exists out there). It also fixes a couple of bugs in the fakexa
implementation revealed by using offscreen pixmaps.
2006-03-10 21:36:24 +00:00
5b1a7b478f
Move the exaDrawableDirty in exaPrepareAccess to exaFinishAccess, which is
...
after the drawing is done. Previously, a failed PrepareAccess could
have migrated and cleared the dirty flag before the damage was ever
done.
2006-03-10 21:32:34 +00:00
ffdbb547be
Coverity #1011 : Remove a useless NULL check on a pVbe that had been
...
dereferenced many times before.
2006-03-10 08:06:42 +00:00
1bc72dce5f
Coverity #857 : Fix resource leak in error path by freeing earlier.
2006-03-10 08:03:24 +00:00
55f677d600
Coverity #813 , #814 , #815 , #816 : Fix resource leaks in error paths of
...
config parsing code.
2006-03-10 07:58:27 +00:00
2bd4110549
Document the restriction on PrepareAccess() failure, from discussion with
...
benh.
2006-03-10 07:45:25 +00:00
21dcd03048
Just like FreeBSD, let DragonFly's default mouse Device be /dev/sysmouse
...
(since /dev/mouse don't even exist by default).
2006-03-10 01:34:45 +00:00
9a99afdfb2
Add DragonFly support. (It is like FreeBSD.)
...
This patch is from DragonFly developer Joerg Sonnenberger and the pkgsrc
collection.
I tested using /dev/sysmouse with moused using my serial /dev/cuaa0.
2006-03-10 01:22:26 +00:00
7a0f7f7398
Coverity #349 : Fall back to software early if pSrc->pDrawable is NULL, or
...
pMask is non-NULL but pMask->pDrawable is NULL. This prevents NULL
dereferences on gradients and other Pictures which have no pDrawable.
2006-03-09 23:29:44 +00:00
ab35c3fbc1
Do a first pass of doxygen documentation of EXA. This removes the
...
corresponding pieces of exa-driver.txt, which were becoming stale.
Hopefully the documentation will stay much more up-to-date this way.
Many thanks to jbarnes for writing exa-driver.txt which was used a lot
in writing this documentation.
2006-03-09 23:18:15 +00:00
d8f8bfecce
Do Xorg configure checks for Xgl only as well
2006-03-09 14:23:57 +00:00
2822cbc1fb
Rearrange EXA driver structures so that there's a hope of maintaining ABI
...
when extending the driver interface. The card and accel structures are
merged into the ExaDriverRec, which is to be allocated using
exaDriverAlloc(). The driver structure also grows exa_major and
exa_minor, which drivers fill in and have checked by EXA
(double-checking that the driver really did check that the EXA version
was correct). Removes exaInitCard(), which is replaced by the driver
filling in the rec by hand, and the exaGetVersion() and related
EXA_*VERSION which are replaced by always using the XFree86 loadable
module versioning.
2006-03-09 06:04:07 +00:00
65aa33f917
render/picture.c Initialize the format of a source picture to
...
PICT_a8r8g8b8. Fixes a failure in the gradients test of rendercheck. In
the long term we could do better by setting the format to something
without alpha whenever the gradient doesn't contain colors with alpha.
This triggers a reduction of the over operation to a pure source
operation.
2006-03-08 06:19:37 +00:00
cb5090e8d6
Bug #6150 : Do the obvious fix of an insane sanity check in
...
xf86InitFBManager. (Julio M. Merino Vidal)
2006-03-08 03:32:07 +00:00
2e6f801fe1
Numerous amounts refactoring and comment adding (see ChangeLog for file by
...
file details). The primary intention for these changes is to pave the
way for the new device probing and PCI configuration code that I'm
working on.
2006-03-07 23:58:22 +00:00
b7d2dfc1e5
Add appropriate MIT license. Oops.
2006-03-07 20:06:15 +00:00
9d8c0e4bcb
Add a new flag to ephyr, "-fakexa", which turns on an EXA acceleration
...
implementation that calls fb to get its work done. The purpose is to
have a trusted EXA driver for use with testing changes to the core of
EXA. However, fakexa has not received much testing yet, lacks offscreen
pixmaps support, and doesn't reliably provide garbage when EXA doesn't
get its syncing right. All of these should be fixed soon.
2006-03-07 19:57:46 +00:00
0a3d6c7399
Remove stale EXA files, which failed to get removed during the move to
...
top-level, somehow.
2006-03-07 19:49:31 +00:00
68a8963f72
Fix cvt -r check again. CH7011 TV encoder had 800x600 PAL hit the check.
2006-03-07 16:00:57 +00:00
0693083335
render/picture.c Correctly initialize devPrivates variable in source only
...
pictures to 0
miext/cw/cw.h Don't try to access devPrivates of source only pictures
2006-03-06 21:00:09 +00:00
448997ebcd
Only output SetClientVersion message if verbosity > 1, like other
...
extensions do
2006-03-05 16:43:10 +00:00
d921173833
define SYS_LIBS to hold system dependant libraries that may needed. and add
...
it to libraries list where needed. Update ChangeLog for previous
changes too
2006-03-05 16:35:08 +00:00
4335868476
Fix build with non GNU make.
2006-03-05 16:13:21 +00:00
96ca329382
Remove redundant definition of struct Inst. Safeguard xf86AddDriver against
...
future additions to DriverRec.
2006-02-28 23:07:09 +00:00
1cfa9f647e
Bug #5216 : Allow options to appear with other components.
2006-02-28 16:55:26 +00:00
e3b6b95f29
Bug #5627 : Fix Xprint font symlinking. (TIlman Sauerbeck)
2006-02-28 16:26:16 +00:00
345d99c972
Typo fixes (Nicholas Joly, XFree86 bugzilla #1658 )
2006-02-27 16:19:39 +00:00
f41ec003f3
Add two radeon pci ids (one is used for a radeon mobility X700 XL in a
...
medion notebook, the other is for a AIW X800 VE)
2006-02-25 01:17:10 +00:00
01a0bf881a
OpenBSD supports PCVT and WSCONS. PCCONS is long gone.
2006-02-24 17:01:57 +00:00
5d9a620726
Remove redundant composite op reduction, done in Render now.
2006-02-23 19:25:57 +00:00
028d6903f6
Bug #5978 : Added missing swap of input variables. Added missing cases for
...
GL_SECONDARY_COLOR_ARRAY and GL_FOG_COORD_ARRAY (Colin McDonald)
2006-02-22 16:29:07 +00:00
43324132af
Added Mesa include directory
2006-02-22 16:08:56 +00:00
43fbcc28c9
Update to build against Mesa HEAD. (Thank you Felix for feedback on xorg
...
list.)
2006-02-22 02:47:00 +00:00
cfdacab33a
Eliminate unused PCI BIOS reading functionality. The old code used several
...
function pointers to implement a level of flexability that was never
used. The code also had unused support for extracting a single image
type from a larger expansion ROM.
Fix the spelling of PCI_BIOS_OPEN_FIRMWARE.
Fix a couple errors in #ifdef debug code.
These changes have been tested on x86 and x86-64 Linux.
2006-02-21 00:02:08 +00:00
5fd0f94006
Eliminate all the code for querying the PCI class from the PCI ID database.
...
Class information is not, and never has been, stored there. Therefore,
this is just a bunch of elaborate code to read 0x00000000.
This has received testing on x86 and x86-64 Linux.
2006-02-20 23:45:50 +00:00
6d7083bd69
indent fixes (OMG SO UGLY), and nuke old RCS keywords.
2006-02-20 22:16:49 +00:00
5480c537ce
ANSIfy Xi/. Mostly automated via protoize(1).
2006-02-20 21:50:49 +00:00
ed3ea887a6
Fix the encoding of DES's name.
2006-02-16 21:49:51 +00:00
ea0b3f65f7
Fix build of Xorg by putting xf86bigfont back into builtin instead of
...
module sources list.
2006-02-16 21:45:12 +00:00
d97f29be22
Bug #5453 : Don't forget to still do AM_CONDITIONAL for XVMC even if XV is
...
disabled, and also force XVMC disabled if XV is disabled. (Dag-Erling
Sm�rgrav)
2006-02-16 21:35:32 +00:00
60d4839a2d
Forgot to include this file in previous patch
2006-02-16 19:03:07 +00:00
8987b2c1ef
Make more extensions optional in build (for kdrive). Fix kdrive build for
...
actual hardware. Fix kdrive pointer signed/unsigned types. Add
kdrive-required YX rotation functions. Replace rgb text file loading
with static rgb color table.
2006-02-16 07:17:31 +00:00
c8acb34269
Fix an occasional crash on VT switches: the server would save a pointer to
...
the current cursor when disabling FB access and would try to restore
that cursor when re-enabling. However, that cursor might have been
destroyed in between. This fixes it by updating the saved cursor
pointer when a cursor is set and vtSema is FALSE.
2006-02-16 06:52:12 +00:00
c845e152f5
Bug #5871 : Drop special build infrastructure left over from libcwrapper
...
times. (George Fufutos)
2006-02-16 05:46:08 +00:00
c2f685e644
Fix XSERVER64 should be _XSERVER64
2006-02-16 05:18:20 +00:00
1bbd5e49b1
Bug #5869 : Remove traces of EXTMODULE define, which doesn't appear to be
...
useful any more. (George Fufutos)
2006-02-16 04:59:45 +00:00
c03b06bdf0
Bug #5888 : Remove orphaned laymodule.c from miext/layer removal. (George
...
Fufutos)
2006-02-16 04:39:00 +00:00
3b32e902c7
Add entry missed in my last commit.
2006-02-16 01:49:23 +00:00
c4767794ef
Export xf86ActivateDevice, used by the evdev driver.
2006-02-16 01:03:09 +00:00
6770f1bdb1
Define NO_LIBCWRAPPER in dix-config.h, and rely on Mesa including
...
dix-config.h if DIX_HAVE_CONFIG_H is defined to get it and _XSERVER64,
instead of defining things like this per directory.
2006-02-15 21:09:14 +00:00
7d7fc927cd
Remove a few #ifdef vms; whatever problem that was solving should assuredly
...
be solved some other way.
2006-02-15 20:47:44 +00:00
f105b8da11
Mark everything in dixsym.c as _X_EXPORT.
2006-02-15 20:44:13 +00:00
010d6effa6
Mark everything in {ext,font}sym.c as _X_EXPORT.
2006-02-15 19:15:32 +00:00
3fe482c77e
Mark everything in misym.c as _X_EXPORT.
2006-02-15 19:05:55 +00:00
6ad4325b87
Update to build against mesa head.
2006-02-15 18:26:45 +00:00
50e2ff9a25
Remove the waitSync from KdDisableScreen and push it off to drivers'
...
disableAccel hook, which is more correct anyway. This makes kdrive.c
not have any knowledge of kaa, opening the way for using exa from
kdrive.
2006-02-15 03:20:55 +00:00
0446aafa94
Avoid some more libcwrapper damage that prevented kdrive linking.
2006-02-15 03:07:23 +00:00
5c9b6f0fb0
DRIGetSecs() would call getsecs() when XFree86LOADER is defined, relying on
...
the wrappers to provide it. Wrapper gone, and getsecs doesn't exist on
linux so it now blows up. Fixes it by just calling gettimeofday() in
all cases instead.
2006-02-14 08:14:42 +00:00
049dca0f43
Remove useless line of code that contained a bug and triggered a gcc
...
warning. This variable will be overriden before being used anyway.
(Bugzilla #5595 )
2006-02-14 08:11:41 +00:00
bb8c36690a
Bump shadow module version number to 1.1.0 from 1.0.0
2006-02-14 04:20:37 +00:00
cc9dfab0b3
(Reverted)
2006-02-13 18:57:38 +00:00
4a7f6f53ca
Further op reduction when both src and dst alpha are absent.
2006-02-13 18:09:51 +00:00
28ced9f3e0
Add missing ChangeLog text for r1.2 of GL/glx/indirect_reqsize.h
2006-02-13 05:29:00 +00:00
4839e91fca
HAS_MKSTEMP vs. HAVE_MKSTEMP (From Fredrik Höglund)
2006-02-13 05:03:13 +00:00
1a4f20541a
Fix linux build without libc wrappers (From Fredrik Höglund)
2006-02-13 04:56:27 +00:00
2dc7b5e0d9
Move call to xf86WrapperInit() to OsVendorInit() in xf86Init.c and remove
...
stubs in other DDX.
2006-02-13 04:43:40 +00:00
34d0b9228f
Simplify ops that would use the alpha channel when an alpha channel is
...
always 1.0, and short circuit PictOpDst for good measure.
2006-02-12 20:53:35 +00:00
Eric Anholt
Daniel Stone
Adam Jackson
Matthieu Herrb
Felix Kuehling
Benjamin Herrenschmidt
Kristian Høgsberg
Deron Johnson
Alan Coopersmith
Jeremy C. Reed
Matthias Hopf
Lars Knoll
Ian Romanick
Luc Verhaegen
Roland Scheidegger
Alexander Gottwald
Keith Packard
Dave Airlie
Zephaniah E. Hull
Alan Hourihane