Changes:
Save/restore ECK Clock Synthesizer
Fixed bug with panel autodetection using
ViaPanelGetNativeModeFromScratchPad function
Fixed problem with backlite (Ticket #308)
TV-out VT1625 chip support
Cursor support speedup
Removed two old variables:
pBIOSInfo->panelX
pBIOSInfo->panelY
and replaced it by:
pBIOSInfo->Panel->NativeMode->Width
pBIOSInfo->Panel->NativeMode->Height
Fix bug with DFP Power Off (ticket #317)
Fix bug with wrong panel size value
(http://www.openchrome.org/trac/ticket/301)
This bug appeared only when UseLegacyModeSwitch was TRUE
Added support of the VIA OpenBook
Added support for screen rotate upside-down and remove "magic numbers"
Forgot to remove an old Dot Clock entry from the table.
XO 1.5 Panel patch contributed by Xavier Bachelot.
Fixup some of the Dotclock code and add working plls
for the XO 1.5
CX700 integrated TMDS (Patch courtesy Tim Chen from VIA).
Initialize CRTC before a mode switch. Fix bug #260.
a couple of small compatibility fixes for XAA and EXA.
viaLowLevel.c and via_accel.c are fixes by
Bartosz Kosiorek. The via_video.c is a syntax
fix by Robert Bridge. Thanks for the bug fixes.
Patch submitted by Robert Bridge, originally started by
Xavier Bachelot, to remove old EXA support. We are moving
forward with the driver and are no longer supporting the
old EXA initializations.
According to VIA's code the P4M900 should have the same
acceleration initialization registers as the K8M890. This
patch fixes this and gives about a 10x increase in the
video benchmark numbers.
remove the check for VIA_3D_ENG_BUSY in the idle loop.
This causes xorg to chew a bit more cpu with XAA enabled,
and it causes the entire X server to dump if EXA is enabled.
I should investigate further, but for now I will just remove
it.
Oops forgot to change registers for the vx800 chipset
when waiting for Idle in the engines. I also added
back in waiting for the 3d engine to be idle. I am
not sure if this was removed for a reason.
Temporary hack to fix a bug with newer Xorg and DRI
that crashes if CRTC is not setup. I am adding this
because I am working on this part of the code anyways
so it won't get lost in bit rot.
Make sure that the monitor settings will handle our
internal Modes set for TV out and LCD Panels.
A couple more tweaks to fix mono hardware cursor on older
chipsets.
Missed a couple of changes for the chipset specific
cursors. Moved chipset detection to HWInit. Removed
some extraneous debugging.
Initial XVideo support for VX800.
Okay I tried to cheat. We need to have some chipset
specific intelligence. I have added this back into
the initial test to check for HWCursor support.
This is a take at reworking HWCursor. It works off the
ttm_branch and the work done in the randr_branch. It only needs
marginal testing as HWCursor was already very broken.
Moved pVia->FrameBufferBase setting to within
VIAMapFB where it was before being called without
being set.
Patch submitted by Robert Bridge to clean up compiler
warnings. Should close trac ticket #243. Thanks
RobbieAB.
Finally tracked down the XAA libpciaccess bug. Also
Realized I had not enabled write-combining on the main
framebuffer. This should give a small speed boost.
Added support for all Modes in ViaPanelNativeModes to
ViaPanelModes. Modes used are from Via's released driver
in ViaSupportModes.
Found needed register writes that are needed so VX800 doesn't
crash when CRT is not enabled. I also removed NEEDS FIXING
comments from them as they are obviously needed.
A few more acceleration tweaks for the VX800.
This should fix all the XAA rendering issues.
This should also allow 800x480 Panels to work without
custom modelines.
Initial support for the VX800 chipset. This is based off the
Basic chipset support work of Xavier Bachelot and the inital
Acceleration work done by Harald Welte. I verified the changes
based off the most recent VIA codebaase and fixed a couple of
rendering issues.
Still needs more work on EXA ( font rendering glitches ) and
XVideo.
The OpenChrome project is committed to providing and supporting fully free
and Open Source drivers that take full advantage of the hardware
acceleration of VIA chipsets featuring the VIA UniChrome, UniChrome Pro and
Chrome9 integrated graphics processors.
This will probably replace xf86-video-via in the future (probably after the
next openchrome release).
