summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2007-09-17ChangeLog DOES need to be cleaned, but in MAINTIANERCLEANFILES.Martin-Éric Racine
2007-09-12Creating a release of course requires updating the package version...xf86-video-amd-2.7.7.1Martin-Éric Racine
2007-09-12We no longer delete the ChangeLog in the clean target. We will simply ↵Martin-Éric Racine
overwrite it instead.
2007-09-02Erm... actually, the ChangeLog needs ot be removed, since it's generated at ↵xf86-video-amd-2.7.7.0Martin-Éric Racine
release time.
2007-09-02Commit the clean 2.7.7.0 ChangeLog to GIT.Martin-Éric Racine
2007-09-02Added the standard X.org ChangeLog generation target to Makefile.amMartin-Éric Racine
2007-09-02Reverted autogen.sh changes too, since the 'dist' target does what we need.Martin-Éric Racine
2007-09-02Reverted the autotool file generation from the GIT tree.Martin-Éric Racine
2007-09-02Regenerated configure based on new package version.Martin-Éric Racine
2007-09-02Bumped the package version to 2.7.7.0 and calling it a release.Martin-Éric Racine
2007-08-29Updated the ChangeLog.Martin-Eric Racine
2007-08-29Regenerated ChangeLog and autotool files.Martin-Eric Racine
2007-08-02Fix the upload / composite mask raceJordan Crouse
We fix a race with the upload function when using a composite mask by using the exa core function exaGetPixmapFirstPixel, which does the right thing in X 1.4 and newer. For older versions, the fix is to stall the pipeine to ensure the upload is complete, so thats what we do.
2007-07-25Add the all-important if/then loop to avoidMartin-Éric Racine
squashing the ChangeLog if autogen.sh is executed outside of a git tree.
2007-07-13This autogen.sh implements ChangeLog generation from the git commitMartin-Éric Racine
log. It also calls the correct Makefile macro to clean up after maintainer configuration, instead of cleaning up by hand.
2007-07-10Fix "old-school" MSR accessesJordan Crouse
This fixes the "old" way of reading MSRs (through virtual registers) - we had swapped arguments in one of the macros, which does very bad things to the poor systems it was running on.
2007-07-10Change _X_INLINE to just inline - the define probably isn't neededJordan Crouse
for us unless we go to a system that doesn't support it. We'll cross that bridge when we get to it.
2007-07-06An improved autogen.sh scriptMartin-Éric Racine
2007-07-06Remove the DCON sleep codeJordan Crouse
The DCON sleep code was introduced by a previous commit - for now that will only live in the OLPC tree.
2007-07-06Use the right bit depth when doing EXA copies on the GX.Jordan Crouse
This fixes the image corruption problem on OLPC.
2007-07-06Make sure we turn on the video palette when video starts, and turn it offJordan Crouse
when video ends.
2007-07-06Fix LX video downscalingDan Williams
The GX video downscaling logic apparently isn't appropriate for the LX
2007-07-06uint32_t -> CARD32Adam Jackson
2007-07-06Multiple fixes for the LX driverJordan Crouse
* Fix VGA detection * Restrict EXA implementations older then 2.0 * Fix crash when VT is switched while rotated * Turn the shadow framebuffer into a true exa offscreen component * Shut off video when we are rotated Conflicts: src/amd_lx_driver.c src/amd_lx_rotate.c
2007-07-06Two fixes for compositing for Cairo 1.4.4 and X server 1.3+Jordan Crouse
2007-07-06From Bernardo Innocenti.Zephaniah E. Hull
- enable a few GCC warnings - make a few globals const and static - move MGP_RASTER_MODE writes before writes to color registers - compute BPP dynamically in the SolidFill EXA hook Conflicts: configure.ac
2007-07-06Fix A8 masksJordan Crouse
We were using the wrong operation for A8 masks, resulting in badness. Also, clean up the mask blt to be much simpler.
2007-07-06Use the correct pitch for composite mask operationsJordan Crouse
We need to specify the correct pitch for composite mask operations - this fixes the stride problem. Also, correctly set up the source color and some other minor issues, this helps us pass caps-join in the cairo test suite.
2007-07-06Swizzle the sync when the panel is enabled to account for Cimarron beingJordan Crouse
silly.
2007-05-31Add DCON detection to LX.Adam Jackson
2007-05-31DCON detection cleanup and generification.Adam Jackson
2007-05-31Un-libcwrap.Adam Jackson
2007-05-31Add extra GCC warning fu. -Wall cleanup.Adam Jackson
2007-05-31Rename: amd_gx_dcon.c -> amd_dcon.cAdam Jackson
2007-05-16Remove 'man' directory references from configure.ac.Adam Jackson
2007-05-10Merge Geode GX and LX development from OLPCJordan Crouse
Conflicts: man/amd.man
2007-05-10Documentation cleanupsJordan Crouse
Rework the readme, and delete the now woefully out of date man pages. Also update the TODO file with more interesting information.
2007-05-09Fix warnings discovered by Dan Williams and -WallJordan Crouse
2007-05-09More -Wall warnings fixedDan Williams
2007-05-09Eliminate the first round of warnings discovered with -WallDan Williams
2007-05-04Completely eliminate the rotation pitchJordan Crouse
Chris Ball is still getting segfalts - completely eliminate the call.
2007-05-04Fix a segfault in RandRJordan Crouse
2007-05-04Use the same stride for all rotationsJordan Crouse
This avoids problems copying pixmaps from EXA space.
2007-05-04Cimarron wants to use CF8/CFCJordan Crouse
Avoid using the PCI access routines in Cimarron
2007-05-04Handle out of memory situations more gracefullyJordan Crouse
2007-05-03commit 5f7979e990c5eafb44aa0fa9779a7ed3904d25ceDan Williams
include string.h for memset
2007-05-03commit 7125fcab372b24f7045bbdaa6d4f6e435b1f7a83Dan Williams
Fix missed GXQueryImageAttributes->GeodeQueryImageAttributes
2007-05-02Solve several vexing issues with RandR:Jordan Crouse
* Incorrect rendering while rotated * Mouse was incorrectly drawn for 90 and 270 * subsequent RandR commands were killing the rotated stride
2007-05-01Fix video and VGA bugsJordan Crouse
Fix two small bugs: The VGA dection code wasn't working, and the video engine was not using color keying.
2007-04-26Use a secret bit to ensure we use channel A alpha *before* color conversionJordan Crouse