implementation. Includes cache flushing to prevent bad first reads of
the framebuffer, fixing blending of many formats, falling back on many
unsupported operations, and falling back early to prevent migration.
Passes all of rendercheck except some of the blend (!) tests.
in the kaa structure: markSync and waitMarker. The first, if set,
returns a hardware-dependent marker number which can then be waited for
with waitMarker. If markSync is absent (which is the case on all
drivers currently), waitMarker must wait for idle on any given marker
number. The intention is to allow for more parallelism when we get
downloading from framebuffer, or more fine-grained idling.
- Replace the KdMarkSync/KdCheckSync functions with kaaMarkSync and
kaaWaitSync. These will need to be refined when KAA starts being smart
about using them. Merge kpict.c into kasync.c since kasyn.c has all the
rest of these fallback funcs.
- Restructure all drivers to initialize a KaaInfo structure by hand rather
than statically in dubious order.
- Whack the i810 driver into shape in hopes that it'll work after this
change (it certainly wouldn't have before this). Doesn't support my
i845 though.
- Make a new KXV helper to avoid duplicated code to fill the region with
the necessary color key. Use it in i810 and mach64 (tested).
- support for mouse-like devices that report absolute coordinates
- support for mouse-like devices that need to know screen size so they can
scale absolute coordinates to screen size
- fix 3-button emulation
DRIINFO version to match. Fix various pointer printfs in libdri to be
64-bit aware. Silence a warning about redefining noXFree86DRIExtension.
(Jesse Barnes)
number of the DRIInfoRec, disambiguating it from the XF86DRI protocol
version number. Modify DRIQueryVersion to return the libdri version
number, which all DDXes were requesting implicitly. Fix the DDXes to
check for the DRIINFO version they were compiled against.
standard mode solely determined by the mode ID or it may be a detailed
mode with almost arbitrary mode parameters. The mode parameters are
passed to the BIOS in a memory area pointed to by es:di. If bit 11 in
the video mode is set this memory area is evaluated. When we don't have
such an area (passed to VBESetVBEMode() this function should unset this
bit for sanity reasons. (Bugzilla #3329)
number of preallocated slots. We should really make this dynamic - but
I don't think this ever caused a problem so it's more or less academic.
A. Avoid that *SyncStart starts before *BlankStart. If *BlankStart >
*SyncStart it is made = *SyncStart and its width is made maximal but such
that the blank does not exceed *Total. Since the Sync width has the
same restrictions as the Blank width monitors should still be able to
clamp after the sync pulse. B. Over time mode validation has become
inconsistent when people started to add additional features to the mode
validation. One such feature is that the mode->Crtc* values have been
(ab)used to allow the driver ValidMode() function to pass driver
normalized timing values back to the validation function. The
introduction of these features made the code less readable and created
numerous possibly unintended side effects in the validation semantics.
I've attempted to consolidate these changes making the code more
consistent and eliminating a number of side effects. This should not
cause problems for the majority of drivers, still it should receive
testing - especially with ATi Mach64 and Radeon code. (Bugzilla #3325).
change "foo.h" to <X11/foo.h> for core headers, e.g. X.h, Xpoll.h;
change "foo.h", "extensions/foo.h" and "X11/foo.h" to
<X11/extensions/foo.h> for extension headers, e.g. Xv.h;
change "foo.[ch]" to <X11/Xtrans/foo.[ch]> for Xtrans files.
(https://bugs.freedesktop.org/attachment.cgi?id=2434) Fix Solaris build
bustage caused by namespace collision between symbols defined in
<sys/kbd.h> and those in "xf86_OSlib.h". Patch by Alan Coopersmith
<alan.coopersmith@sun.com>.
Linux-only fixes: Fix case where a smaller write-combining region
blocks write-combining setting of the whole frame buffer. Fix bug in wc
setting code when regions are first splitted and setting of
write-combining then fails.
DriverFunc() also returns FALSE when the specific sub function isn't
supported. In the case of xf86RandRGetInfo() we simply rely on what has
been set before and return TRUE. In the case of xf86RandRSetConfig() we
only bail with FALSE if we have to do a rotation and the call fails. We
presently cannot do rotation on the fly without the help of a driver
function (Bugzilla #2745).
from within winDestroyWindowsWindow
DestroyWindow send a WM_WINDOWPOSCHANGED to another window causing a
restacking of all windows, even of the window which is just about to
destroyed and whose structures may not be intact anymore.
xc/programs/Xserver/hw/xnest/Imakefile
//bugs.freedesktop.org/show_bug.cgi?id=2653) attachment #2020
(https://bugs.freedesktop.org/attachment.cgi?id=2020): Cleanup Xnest
usage of the DPMS dummy stub functions from dpmsstubs.c instead of
using it's own copy of these functions.
(https://bugs.freedesktop.org/attachment.cgi?id=2019): Fix
|xf86SignalHandler()| which resets the signal handler before setting
the flag indicating a signal has been caught, theoretically allowing
the possibility of infinite recursion. Patch by Andrew Church
<xzilla@achurch.org>.
xc/programs/Xserver/hw/xnest/Window.c
//bugs.freedesktop.org/show_bug.cgi?id=2546) attachment #2005
(https://bugs.freedesktop.org/attachment.cgi?id=2005): Fix Xnest to
update the shape regions in the backend server whenever a client
changes them in Xnest (the fix is to add a new wrapper which calls
|xnestShapeWindow()| before calling |miSetShape()|). Patch by Mark
McLoughlin <mark@skynet.ie>
the async windowmanager thread. Fixes some restacking problems occuring
which were timing dependent Do not raise the window on WM_ACTIVATE
Removed unused code for WM_WINDOWPOSCHANGING ESC is debug key. Print
status but do not abort processing the message
winfont.c
winmultiwindowshape.c
winmultiwindowwindow.c
winpfbdd.c
winshaddd.c
winshadddnl.c
winshadgdi.c
Fix incorrect wrapping of functions. Ensure the pointers from pScreen point
to the called function even if wrapped functions changed it
Set the window properties to NULL to avoid referencing freed memory because
of timing problems after deleting a window
Do not wrap ChangeWindowAttributes. All functions are noops currently
When COMPOSITE is enabled, call CopyWindow even when the pixels "don't
move" on screen.
Don't reject modes that are not supported by the unused monitor.
xc/programs/Xserver/hw/xfree86/os-support/linux/lnx_io.c
//bugs.freedesktop.org/show_bug.cgi?id=825) attachment #956
(https://bugs.freedesktop.org/attachment.cgi?id=956): Fix build
problems on Linux/SPARC. Patch by Jeremy Huddleston
<eradicator@gentoo.org>. Approved in the 2005-01-31 Xorg
release-wranglers phone call.
permissions when no access to HW registers is required. For API changes
which mostly involve the modifications to make the RRFunc (introduced
with 6.8) more flexible please check Bugzilla #2407. NOTE: This patch
applies changes to OS specific files for other OSes which I cannot
test.
but includes debugging to ensure that the reg being submitted is the
one that follows in the packet. Convert most uses of OUT_RING to it,
and convert a couple of OUT_REG sets to DMA_PACKET0/OUT_RING_REG. Also,
add checking to see if more registers are submitted to a DMA_PACKET0
than should be, to avoid hangs during stupid mistakes (checking for
less isn't done).
virtually impossible to determine if the old code did what it was
supposed to do. Also changed paradigm: Trust the based address more
than the (guessed) size. (Bugzilla #2383)
reading a register/attempting DMA. Now it'll wait for a certain number
of seconds rather than a certain number of times through the loop
before deciding that it's timed out and resetting the hardware. Also,
add more timeout handling, and reset the draw state after resetting the
engine.
problems on may systems which don't support APM standby and might not
be required as according to the APM specs the chips should receive
enough power to retain its state.
- Print out power state change requests to log file in all verbosity
levels.
- Don't change server state if no driver PM function is registered.
(Bugzilla #2279)
relied on the generic VGA layer to restore text console fonts for us
when shutting down the server or VT switching back to the text console.
This has worked rather well but it has some downsides on Linux: a. Many
people use fbdev as console text mode. In this case it is not necessary
to save/restore console fonts as the console is running in graphics
mode anyway. b. Some architectures don't have a fbdev console but
require a full POST of even the primary card (ie. IA64). This posting
has to take place before we even have a chance to save anything.
Therefore the fonts we save are the once written to the chip by POST,
not what has been programmed by the user. c. Certain chipsets utilize
the BIOS to perform mode setting. This may interfer with the vga
save/restore font function in a strange way. It would therefore be
preferrable to let the OS - which has been used to set up the font in
the first place - take care of saving/restoring the data. I will attach
a patch which will do so for Linux. To make this fully functional a
small patch needs to be applied to the Linux kernel. To disable this
feature add: #define DoOSFontRestore NO to your host.def. (Bugzilla
#2277)
InitOutput.c
XWin.rc
winerror.c
wintrayicon.c
winvideo.c
winshaddd.c
Set PROJECT_NAME in Imakefile to create alternative window titles for
Cygwin/X and Xming
are always emitted. Fix the real problem, which was not enough regs
being initialized in ati_draw.c. Fix a typo that was resulting in alpha
coming out as 0 * src or 0 * broken instead of src * 1 or src * mask.
Assign the blending results to R0, as appears to be necessary. Unbreak
the dst-alpha-blend-with-no-dst-alpha code. Yow. And set the right DMA
count for the r200 traps code.
non-mask rendering. Reenable it. Also, R200TexFormats was used instead
of R100 in one place. Harmless so far, because the formats were in the
same order.
(https://bugs.freedesktop.org/attachment.cgi?id=1530): Fix the current
implementation to make it possible to slow down the mouse pointer or
use arbitrary fractions (without running into rounding error issues).
The change is using the same method of preserving rounding errors that
the exponential method is already using. Patch by Jan Brunner
<Jan_B@gmx.ch>.
needed when building this file
- bug fix: pointerMsg -> keyboardMsg in auto-configuration code.
- make the 'kbd' driver the default for autoconfiguration everywhere.
compiled in or not.
Removed stale make variable.
Updated xaaWrapper.c: Under certain circumstances the accel path was not
used when possible. Removed some debugging stuff and stale code that
had been commented out.
Check if maps are really installed before attempting to list them.
obj_base.h is not needed anymore. Using <objbase.h> instead.
Use Xwindows.h instead of windows.h
do not include win_ms.h
remove extra definition of sleep()
Set HOME to Documents and Settings/username if not set
Use Xming basedir instead of ProjectRoot for system.XWinrc
Fix callback functions to use wBOOL instead of BOOL
Fix compiler warnings. Added debug output.
Fix warning about undefined macro max
Avoid PIO access on IA64. Some IA64 machine check if legacy ports outside
the VGA range are accessed. The ATi driver however does this to probe
for ISA Mach8/32/64. Since no IA64 has ISA slots this restriction
should not be relevant to the user.
Avoid recursive calls of xf86scanpci(). This function normally detects that
it has been called before by checking if the PCI structure is filled
out. So far if this was not the case (because PCI probing has failed
for some reason) the function is traversed again. With the chipset
specific PCI bus probing this can lead to an endless recursive loop as
the post-probing code calls xf86scanpci() from within this function.
The OS specific PCI code for Linux worked only if bus 0 was populated as it
checked for the presence of /proc/bus/pci/00. Fixed to check for
/proc/bus/pci/<bus_to_look_for> instead.
Remove old obsolete include/extensions/damage.h.
Include srvrv_ctrl(xfree86) in macintosh US XKB symbol file so VT switching
works again on mac (#1872).
Remove out of place #define's of printf to xf86printf. This definition
should only be in xfree86/os-support/xf86_libc.h
update references to xf86site.def in comments (Bug #1827).
fix kbd driver for wskbd protocol and pure wscons console driver (Bug
#1825).
don't add '-4' to generated default file name (bug #1826).
typo in resource name (XFree86 bug #1300, X.org bug #1825)
Updating to EDID 1.3. (Bugzilla# 1490, Jay Cotton, Egbert Eich).
Removing unneeded code.
Fixed KGA handling for i810. KGA handling for chips derived from C&T chips
is slightly different. The changes make the code consistent with the
C&T (chips) and i740 drivers.
XScreenSaverSetAttributes() followed by XScreenSaverUnsetAttributes().
Caused by missing FreeResource() in XScreenSaverUnsetAttributes().
Removing unused DDC sections that caused misinterpretation of DDC data due
to a missing break statement in a switch.
Fixed typo: #if <-> #ifdef.
lists themselves. To use with the Sun compilers, add to host.def: #
define UseCCMakeDepend YES # define DependFlags -cc $(CC) -d -xM (Sun
bug id #4245688 - fix by Alan Coopersmith)
Add Solaris to the platforms on which mprotect is run to set execute
permissions when necessary. (Sun bug id #6175128 - fix by Alan
Coopersmith)
Internationalize digital output (Sun bug id #4119396 - fix by Steve
Swales), add -bgpixmap option to set XPM file as background (originally
from STSF project version of xclock by Alan Coopersmith)
xc/programs/xmodmap/handle.c,pf.c xmodmap was printing line numbers which
are one too low in error messages (Xorg bugzilla #1739, Sun bug id
4637857 - fix by Sam Lau)
Added support for IBM space saver keyboard (Stefan Dirsch). Added support
for Cherry CyMotion Master XPress (Marcus Schaefer).
Change order of SetDisplayDevice(), HWRestore(), UnbindGART() and
RestoreBIOSMemSize() to be exactly opposite to the Save procedure in
EnterVT() (Matthias Hopf, Alan Hourihane).
Fix text mode restauration by removing the assumption that the register
which determines which head is programmed is set. to the active head by
the BIOS (Mark Vojkovich).
When I wrote the resource code 5 years ago I made some assumptions which
turned out to be false: I've assumed that the bus number of the PCI
hostbridge would be the PCI bus the bridge links to. This is not
correct. Fixing this assumption is not easy. However I hope that the
attached patch will make the system work regardless as it 'ignores'
host bridges from which the target bus is not known. This should not
matter at all as we really don't care about host bridges (unless we
have bridge specific code which retrieves information about the
bridge).
Fixed server crash on reset when a structure allocated in PreInit() was
freed on CloseScreen().
Fixed ring buffer lock ups that happened because the structure that
contained ringbuffer data was not zeroed after allocation.
Fixed numerous warnings due to signed unsigned comparisons.
programs/Xserver/hw/xfree86/drivers/nv/nv_driver.c:
(NVBacklightEnable): Changed the order in which the sequencer registers
and the backlight control registers are written. The sequencer control
register need to be written first otherwise DPMS screen blanking
produces vertical bars on a mobile device. lib/Xau/Imakefile: Build
libXau static library PIC so it can be linked into toolkits that would
like to wrap its functionality.
version called the driver directly and too late.
Unblank secondary screen explicitely. Don't rely on the value read during
register save as the BIOS have blanked the secondary head.
Checking if server isn't switched away before calling sync. Sanity check
for possible bugs in aother areas of the code.
Fixing default amount of of allocated video memory from AGP for i810: Use
16MB if less than 192MB are installed else use 24MB (Matthias Hopf).
xc/programs/Xserver/hw/xfree86/drivers/i2c/fi1236.c Make sure formatting
style is consistent within a single function. MT2032 functions are best
be in separate file anyway.
code is practically untouched since xatitv (which was a standalone test
program). However, it all worked and was debugged over long period of
time, so I prefer to to mess with these for now.
New drivers: fi12xx (including MT2032 - this would be be split off later).
tdaXXX msp34xx bt8xx
#defines/-D options.
programs/Xserver/hw/xfree86/common/xf86DPMS.c Use defaultDPMSEnabled global
for the default state of DPMS if not set in any config files.
programs/Xserver/hw/xfree86/os-support/sunos/solaris-sparcv8plus.S Add
support for required assembly inline functions for Sun compilers on
Solaris/sparc.
Add support for Solaris/sparc libraries.
Alan Coopersmith from release 6.8.1.
Fail during initialization with error if font/fontset is not set for
widget. This prevents a sig11 later when the non-existent font/fontset
structs are referenced.
Check if xf86Info.kbdProc pointer is really set before calling it on abort
as this pointer won't be set if the new modular keyboard driver is used
(Matthias Hopf).
Added new libs to the bindist control files.
Removed inclusion of unnecessary kernel header on Linux. This may fail in
an -ansi environment.
chipsets can be probed without OS support as probing them is only
possible using ACPI. One example of this are the HP ZX1/2 chipsets:
previously the code assumed that these chips were present when no other
of the known chipsets could be probed. This assumption brought SGI
Altrix machines with 64 CPUs to a grinding halt.
troubles.
- Move the RemoveBlockAndWakeupHandlers to match
RegisterBlockAndWakeupHandlers.
- Enable R100 trapezoid "acceleration" when DRI is working, so that it can
be exposed and worked on.
instead of ErrorFs for things that are really bad, and put limits on
some loops. Now, sometimes instead of hanging the entire system, we
(mostly-) cleanly drop to console when the card has hung.
not doing an add of 1 per triangle like I hoped, and instead seems to
be saturating all the pixels or something. The R100 acceleration
renders pretty well, with some gaps. Note that both are slower than
software due to lack of DMA to submit vertices.
- Mostly fix R128 and Radeon transform support, including supporting
bilinear filtering on R128. Subpixel offsets are still probably an
issue (reported by rendercheck), but I want to make 100% sure about my
understanding of the protocol before changing everybody, including fb.
- Add support for dst formats without alpha to R128 Composite.
- Remove the R128 Blend code, which has long outlived its usefulness. (I
kept it around for one reason: It could be useful for the w/h > 1024
case with no mask and a non-src op. That seems pretty infrequent and
not worth the trouble).
might be able to do Precise in hardware, so leave it up to the driver.
- Add a helper function for computing a set of offsets for smooth trapezoid
rasterizing using many sharp trapezoids.
RasterizeTrapezoid screen function. These hooks will be called for
imprecise, non-sharp trapezoids with A8 destinations.
Note that the current main consumer of trapezoids, cairo, is requesting
precise, sharp trapezoids by not changing the default Picture
attributes, but gets non-sharp effects in software because fb bases its
choice of sharp/non-sharp on the mask format being A8 vs A1, and cairo
asks for A8. Follow fb's (poor?) example by ignoring the sharp setting
and basing the choice off of the mask being A8.
heuristic, delay the decision until the first
kaaPixmapUse{Screen|Memory}, and put it in framebuffer if UseScreen was
called. Provides a significant improvement in cairo speeds (100%
speedup in cairogears here) and is likely to improve text performance
as well.
Fix build failures when UseDeprecatedKeyboardDriver is YES and
DoLoadableServer is NO (Bug #1229, Kristian Høgsberg).
Fix failure when using DLL loader and LD_BIND_NOW is set (Bug #1212, Adam
Jackson).
None works.
Copy filter to backing picture during validation.
Mark picture serialNumber when setting Filter or Transform so Validate
occurs.
Initialize xf86Screens[i]->pScreen to NULL so that RADEON driver doesn't
crash during server reset using old pScreen.
Fix for lockups on some versions of Matrox Mystique. #687, Patch from Mike
Harris.
Call xf86EnableDisableFBAccess though the function pointer instead of
directly. #1041, Patch from Aaron Plattner.
Swap the phsyical size of the screen when rotiation. #1050, Patch from
Aaron Plattner.
Fri Aug 13 19:47:12 2004 Soeren Sandmann <sandmann@redhat.com>
Make HAVE_FT_BITMAP_SIZE_Y_PPEM conditional on the FreeType version instead
of proping it. This way it will work with the monolithic version too.
#1062, Patch by Owen Taylor.
Xnest was half-using midispcur and doing a bad job of it. Replace all of
that code with mipointer which does a lot of the work.
Support DDXen which don't provide GetWindowPixmap, or which return NULL for
the root pixmap.
Harold L. Hunt II, Alexander Gottwald).
Fix crash when using X core font in zh_CN.UTF-8 locale (Bug #368, Yu Shao,
David Dawes).
Fix glXMakeCurrent(Dpy, None, NULL) crash (Bug #719, Adam Jackson).
HP-PA build fix (Bug #828, Guy Martin, Paul Anderson).
Fix SDK build for GATOS and Wacom driver (Bug #829, Bryan Stine).
Fix attempt to read video ROM before enabling it (Bug #843, Ivan
Kokshaysky, Mike A. Harris).
Fix detection of primary adapter (Bug #843, Ivan Kokshaysky, Mike A.
Harris).
Clarify xset man page description of how to use the keyboard repeat rate
settings (Bug #846, Mike A. Harris).
Fix problem where print-screen key would get remapped to sys-req in certain
keymaps, which broke GNOME printscreen functionality (Bug #847, Owen
Taylor).
Fix several render problems:
- MMIO mode support
- Hang on IGP chips
- VT switching hang
- 3D render corruption (Bug #922, Hui Yu).
the other render symbols.
Initialize screen_x and screen_y to 0. This fixes a problem with the new
Damage based sprite, where intersection test against the root pixmap
would fail because the two fields were not initialized (damage.c,
166-170).
2.1.8 (Chisato Yamauchi).
A small backward compatibility fix to make the freetype module build with
freetype version < 2.1.7.
Fixed/added some debugging code.
SetupForCPUToScreenAlphaTexture2. These add a dstFormat argument after
the previous format argument, which the driver needs to use to properly
set up the destination format. Two new arrays are added for the list of
destination formats supported that correspond to the previous format
arrays for sources.
- Make Render acceleration only occur when the new hook for that
acceleration type is supplied and the dst format list is set, along
with the src format list being set. Without knowing the destination
format, the Render acceleration couldn't properly support all the
destinations it might encounter.
- Bump XAA module minor version.
- Update the Radeon Render acceleration to use the new hooks when the XAA
module is sufficiently new. Fix a bug in the src/dst alpha booleans for
ops, and use them to set blend_cntl to support destinations without
alpha. Add missing PICT_a1r5g5b5 texture format, and add list
terminator. (!)
2004-08-02 Kensuke Matsuzaki
Fix the bug that we can't copy & paste multi-byte string to Unicode-base
Windows application. Rename fUnicodeSupport to fUseUnicode, because it
don't mean wheather Windows support Unicode or not.
XAACachePlanarMonoStippleWeak to XAAGetCachePlanarMonoStipple() for
consistency with other parts of xaa and fix forward declaration in
xaalocal.h. Fixes last comments on Bug #962.
Fixed FreeType module to build with FreeType versions older than 2.1.7.
Fixed typo.
Added vtSema to protect call of driver DPMS function.
removed unneeded variable
Modified RandR driver hook to reduce the number of function calls to one.
Function is sufficiently generic to be extended in the future.
This file currently contains the Cygwin/X specific release notes from 6.7.0
and a first collection of what will make up the release notes for the
next release