summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2009-09-27Geode 2.11.6xf86-video-geode-2.11.6Martin-Éric Racine
2009-09-27Revert "LX: validate display modes"Martin-Éric Racine
This reverts commit 3af472f4453f8e92a9f303e64e2796836374d980. This commit had broken detection of arbitrary non-VESA resolutions, which affected the OLPC XO-1 and a number of Geode-based notebooks.
2009-09-22Geode 2.11.5xf86-video-geode-2.11.5Martin-Éric Racine
2009-09-19LX: validate display modesOtavio Salvador
To avoid using virtual desktop by default we now validate the display modes. Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
2009-09-19GX: refactor code that allow backward compatibility with old video interfacesOtavio Salvador
Most of dependent code is now located in a single place. The macro has been changed to use a cleaner name to be easier to spot why it is needed. Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
2009-09-18geode: further precised brand names of Geode variants by NSC in README.Martin-Éric Racine
2009-09-18Geode: corrected the product names in README.Martin-Éric Racine
2009-09-18geode: Added the PCI vendor and device ID of NSC SC1400 to README.Martin-Éric Racine
2009-09-09Switch project URL in README to read-only www host.Martin-Éric Racine
2009-09-09geode: use lowercase vendor idOtavio Salvador
2009-09-09Added extended info about which Geode driver covers what Geode variant in ↵Martin-Éric Racine
README.
2009-09-05Fixed a typo in src/lx_output.c debug message.Martin-Éric Racine
2009-09-03Added README about X server dependency bump due to GX Randr dixSetPrivate fix.Martin-Éric Racine
2009-08-31Release 2.11.4.1.xf86-video-geode-2.11.4.1Martin-Éric Racine
2009-08-31Fixed typo in GX2 dixSetPrivate patch. Kudos to Eamon Walsh!Martin-Éric Racine
2009-08-28Release 2.11.4.xf86-video-geode-2.11.4Martin-Éric Racine
2009-08-18GX: use XAA by default.Otavio Salvador
EXA support is broken for GX so we disable it for now.
2009-08-18Revert "GX: use XAA by default."Martin-Éric Racine
This reverts commit 09ddb88781bc9ec12531414beabd95e50237fdd5. Otavio just sent me a nicer version of this patch.
2009-08-18GX: use XAA by default.Otavio Salvador
2009-08-18Fix GX RandR to properly use dixSetPrivate.Otavio Salvador
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
2009-08-04Migrated the project URL to the new http://wiki.x.org/wiki/GeodeDriverMartin-Éric Racine
2009-08-04Somewhat harmonized the README between xf86-video-cyrix, xf86-video-nsc and ↵Martin-Éric Racine
xf86-video-geode. TODO: * Rename http://wiki.x.org/wiki/AMDGeodeDriver to http://wiki.x.org/wiki/GeodeDriver and install a redirect from the old page.
2009-07-29Added missing development dependencies (X.org macros) to FAQ.Martin-Éric Racine
2009-07-29Added FAQ about building drivers from GIT for testing purposes.Martin-Éric Racine
2009-07-29Added FAQ about submitting patches and producing useful backtraces.Martin-Éric Racine
2009-07-28geode: change to using ABI version checkDave Airlie
2009-07-28geode: update for resources/RAC API removalDave Airlie
2009-07-28geode: switch to using correct pci config interfaceDave Airlie
2009-07-16Use XORG_CHANGELOG macro to create ChangeLog. #22611Gaetan Nadon
Adding the macro in configure.ac and use it in Makefile.am Refer to: https://bugs.freedesktop.org/show_bug.cgi?id=22611 Global maintenance on all modules Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-07-16Update to xextproto 7.1 support.Peter Hutterer
DPMS header was split into dpms.h (client) and dpmsconst.h (server). Drivers need to include dpmsconst.h if xextproto 7.1 is available. Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
2009-06-19Release 2.11.3.xf86-video-geode-2.11.3Chris Ball
Signed-off-by: Chris Ball <cjb@laptop.org>
2009-06-10geode: add inputstr.h include explicitly.Dave Airlie
2009-05-28Remove useless loader symbol lists.Adam Jackson
2009-05-25Cleaned up the FAQ about WXGA modes.Martin-Éric Racine
2009-05-25Added a FAQ about disabling Panel in BIOS to enable WXGA modes.Martin-Éric Racine
2009-05-23Fix GPIO BAR detection with libpciaccessOtavio Salvador
This patch fixes an issue with the GPIO BAR detection. Basically, with libpciaccess we're finding the ISA device and checking its BARs, but we're not bothering to fill in the BAR information via pci_device_probe. This results in a 0'd out bar and: (EE) GEODE(0): Could not find the GPIO I/O base Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Andres Salomon <dilinger@collabora.co.uk> Signed-off-by: Chris Ball <cjb@laptop.org>
2009-05-12Release 2.11.2.xf86-video-geode-2.11.2Chris Ball
Signed-off-by: Chris Ball <cjb@laptop.org>
2009-05-12Crasher fix: Use ExaDriverAlloc() to calloc the EXA struct.Kyle McMartin
Patch from Kyle McMartin. A recent exa/exa.c commit from Dave Airlie (02ae85c4..) added a new member to the EXA struct, and corresponding code to call it if it's non-NULL. We were using calloc with sizeof() that struct at driver-compile-time; as a result, after the ABI change the new member contained garbage, passed the EXA test for NULL, was called, and segfaulted at startup. RH bug https://bugzilla.redhat.com/show_bug.cgi?id=500086 Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-27Revert EXA 3 build fix.Chris Ball
The EXA developers have backed out the ABI bump to version 3. Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-24Build fix: Include config.h earlierChris Ball
geode.h includes exa.h, without anything having included config.h (which sets the EXA major version) beforehand; add config.h to geode.h. Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-24configure: use AC_DEFINE instead of shell substitutionChris Ball
(This is just cosmetic.) Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-24Build fix: set EXA_DRIVER_KNOWN_MAJOR=3.Chris Ball
We don't use {Prepare,Finish}Access, and: /home/cjb/xorg-build/include/xorg/exa.h:45:2: error: #error Make sure this EXA driver either does not have Prepare/FinishAccess hooks or that they can handle EXA_PREPARE_AUX*, and #define EXA_DRIVER_KNOWN_MAJOR 3 before including exa.h Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-16Release 2.11.1.xf86-video-geode-2.11.1Chris Ball
Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-16Makefile.am: use "git log" instead of "git-log"Chris Ball
"git-log" doesn't work on my Fedora machine, but "git log" should work everywhere. Some distributions choose not to ship all the git aliases. Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-15Change OLPC detection heuristic to one that works on upstream kernelsChris Ball
We've been checking whether we have a DCON/are on an OLPC by looking for files from the "olpc_dcon" kernel module. This module isn't upstream yet, so this only works for the OLPC custom kernel. We might as well check for an OLPC kernel module that *is* upstream, so that we can run on unmodified distributions. This patch moves us to checking for "olpc-ac" instead. Signed-off-by: Chris Ball <cjb@laptop.org>
2009-02-15Fix pMsk=NULL segfault at startup on OLPCChris Ball
Commit 7c278551c79d7a5898c627341f58fad59b7ae20d uses pMsk without testing for non-NULL, and this causes an immediate segfault when running Sugar. Signed-off-by: Chris Ball <cjb@laptop.org>
2009-01-17Fallback in case of mask transforms as well.Mart Raudsepp
Fixes attachment #21519 test case on bug #15700, but not the initial report. Detective work by Michel Dänzer. Signed-off-by: Mart Raudsepp <mart.raudsepp@artecdesign.ee>
2008-12-092.11.0 releasexf86-video-geode-2.11.0Jordan Crouse
2008-12-09xf86-video-geode: after reloading cursors, hide the cursorAndres Salomon
Whenever we reload cursors, the geode driver needs to call hide_cursor() to ensure that the cursor is hidden until we finally unhide the cursor. Without this, we see ghost cursors during rotation or VT switches. Signed-off-by: Andres Salomon <dilinger@debian.org>
2008-12-09xf86-video-geode: fix up cursor size (again)Andres Salomon
The geode hardware (and cimarron) expects cursor memory size to be 48x64, and cimarron will zero out the excess when a smaller cursor is used. It would be nice to be able to use the full cursor size, but xorg's interleaving stuff requires a cursor width that's a multiple of 32. Thus, we're stuck back at 32x32 for cursor size, but we allocate enough memory for a 48x64 cursor. This fixes misc interleaving corruption that we see with 2bpp cursors (when converting to ARGB). Signed-off-by: Andres Salomon <dilinger@debian.org>
generated by cgit v1.2.3 (git 2.25.1) at 2024-12-18 10:36:41 +0000