xserver-xorg-video-openchrome: Changes to 'refs/tags/xserver-xorg-video-openchrome-1%0.3.3-1'

To: debian-x@lists.debian.org
Subject: xserver-xorg-video-openchrome: Changes to 'refs/tags/xserver-xorg-video-openchrome-1%0.3.3-1'
From: Maarten Lankhorst <mlankhorst-guest@alioth.debian.org>
Date: Wed, 25 Sep 2013 14:40:42 +0000
Message-id: <[🔎] E1VOqGg-0000vW-MU@vasks.debian.org>

Tag 'xserver-xorg-video-openchrome-1%0.3.3-1' created by Maarten Lankhorst <maarten.lankhorst@canonical.com> at 2013-09-25 14:40 +0000

Debian release 1:0.3.3-1
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
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=kBS3
-----END PGP SIGNATURE-----

Changes since xserver-xorg-video-openchrome-1_0.2.904+svn1050-1:
Alan Coopersmith (2):
integer overflow in uniDRIOpenConnection() in libchromeXvMC* [CVE-2013-1994 1/2]
integer overflow in uniDRIGetClientDriverName() in libchromeXvMC* [CVE-2013-1994 2/2]

Bartosz Kosiorek (80):
Add rotate upside-down and remove "Magic numbers"
Updated manpage
RandR initial support
Unified variable names with RandR
Remove build warnings and removed unused variables
Add more comment and add basic XVideo support for VX855
Added support for the VIA Openbook laptop (thanks to chfoo)
Add support of the VIA Openbook
Fix bug with wrong panel size (Ticket #301)
For old chipsets (eg. KM400) set UseLegacyModeSwitch = TRUE
Add more traces to help resolve panel resolution bugs.
Fix bug with DFP power off (ticket #317)
Removed two old variables:
Cursor support speedup
TV-OUT, VT1625 chipset support
Fixed problem with backlite (Ticket #308)
Add more traces to solve bugs with panel resolution detection
Fix bug with panel autodetection
Add support duoview for VX855 and add more comment to ViaRestore function
Save/restore ECK Clock Synthesizer
Initial AMD64 support
Add checking of the initialization of the XFree86 framebuffer manager
Add support for Twinhead H12V (P4M900)
Add 1280x800 panel resolution support
Fix panel autodetection for Legacy mode switch
Update openchrome man page, with information about TVPort option.
More debug information to video decoding
Add secondary FIFO support for K8M890 and P4M890, and add more comments
Check the primary FIFO for VX855
Add panel scale support for CLE266 and KM400, fix bug with malloc.
Enabled new mode switch for PM800 chipset,
Add option to enable unaccelerated RandR rotation ("SWRandR").
Enable new mode switch for VM800 chipsets to resolve bugs #240 and #282
More debug traces to via_video.c
Fix manual and add more comments
Fix start address
Optimization of the hardware accelerated cursor and XAA (K8M890 chipset).
Code clean up
Fix bug with suspend and VT switch on VX800 chipset and 64bit systems
Synchronize via_drm.h with drm-2.4.17
Fix segfaults with EXA and XV (Ticket #359)
Fix build on old kernels
Replace RegionsEqual with REGION_EQUAL and use
Manpage update and some minor changes
set colorkey for 2nd_monitor
Fixed freeze on 64bit system for K8M090 chipset
Increase bandwidth to handle 1920x1200 resolution with DDR266
Fix several typo in code comments
Siragon ML-6200 laptop support
update VX855 FIFO
Fix bug #342 with TV out flickering
Replace the deprecated functions with new ones
Replace remaining xalloc to malloc
Change maximum line pitch and virtual height according to chipset
Update 1024x600 modeline
Add workaround for memory autodetection
Enable new mode switch and panel support on K8M800 and VM800
Merge VX900 branch to add basic VX900 support
Fix type in VIASave
Use DRICreatePCIBusID when available to create Bus ID string
Restore video interrupt flag
Fix hardware cursor for VX900
Fix OpenGL application crash on VX900 chipset
Fix VIA VB8001 Mini-ITX Board (P4M900) support
Add suport for Semp Informática Notebook IS 1462 (reported by Colin)
Fix #395 - revert ViaPanelGetSizeFromDDC renaming (thanks to julienvdg)
Add workaround for #177 ticket
Add HP t5550 Thin Client support (thanks to dna67)
Fix resolution detection for OLPC 1.5
Fix Xvideo crash on X.Org server 1.10
Fix XAA displaying issues
Fix cursor garbare after suspend/hibernate/resume on VX855/VX900 chipsets #405
Revert r922 changes (it added I420 support)
Fix cursor garbage after suspen/resume for VX855 and VX900 (#405)
Workaround EXA crash with new libcairo2 (#298)
Initial XVideo support for VX900 chipset
Fix XVideo on VX900 chipset
registers dumper improvements
Fixes to reenable EXA composite support. Currently EXA crashes
Fix Xv bandwidth check when no vertical refresh is available

Benno Schulenberg (73):
move definition of AGP_SIZE out of #ifdef block
comment tweaks
Whitespace adjustment; getting back close to Xorg.
Comment tweaks and #endif clarifications.
Rearrange declarations to fix compilation problem with gcc-2.96.
Textual tweaks.
add an ID, remove a duplicate, remove an old unknown
comment and whitespace tweaks
put in a few extra debugging lines
fix man-page symlink for out-of-tree build
mention the overall revision number in the log
moved a line to after declarations, tweaked comments and debugging lines
Add ability to change gamma setting.
Switch encoding of ChangeLog file to Unicode.
Whitespace and format adjustments of ChangeLog file.
VN800 -> VM800; compile fix
Try CR39 when CR34 says zero during memory detection on the CLE266.
Limit x value to sixteen bits to fix a rendering glitch.
Textual tweaks and order options alphabetically.
Fix typo in man page and neaten changelog.
VN -> VM, message tweaks, and trailing white space fixes
Move reading of the VideoRAM option to after X has parsed the
Print VideoRAM again only when it's actually been probed.
Redoing a man-page tweak.
Removing a duplicated ChangeLog entry.
Cleaning up the ChangeLog:
Redoing the first change of June 10: Putting the chipset and
Redoing the second change of June 10: Printing VideoRAM again
Applying Gabriel's indentation changes (-c439 from randr_branch) also to trunk.
Many more indentation and whitespace fixes.
More whitespace changes.
Whitespace and indentation fixes.
A few tiny changes to bring trunk and randr branch a bit more in line.
Another bit of bringing trunk and randr branch in line.
Whitespace and comment tweaks (in via_cursor.c).
Fix a tiny transposition mistake, and copy the transposition to trunk.
Adding missing comma.
Identifying the driver as Openchrome, with CHROME as a shorthand.
Tweaking and snipping some comments.
Moving memory detection to before the reading of the config file,
Adding more chipset numbers;
Some comment and whitespace tweaks;
Removing unused options; whitespace and message tweaks.
Whitespace corrections.
Whitespace adjustments and comment tweaks.
Moving a little more computation into the helper function.
Tweaking the man-page text, clarifying the effect of an option,
Tweaking a comment, and preferring clarity over saving a few bytes.
Removing the redundant HWCursor option.
Moving the newline in the buildcomment from Makefile to program text.
Correcting print-out of device ID.
Whitespace changes in viaSetAgpMode().
Whitespace adjustments, comment and message tweaks.
Condensing an unneeded code duplication.
Whitespace adjustments and comment tweaks.
Whitespace and comment tweaks.
Whitespace and comment tweaks.
Whitespace adjustments and comment tweaks.
Recreating svnversion.h only when a source file has changed,
Untabbing and a little resort.
Whitespace adjustments.
Whitespace and comment tweaks.
Whitespace and comment tweaks.
Set K8M800 by default to AGP disabled.
Excluding the correct bunch of chipsets from XvMC initialisation.
Replacing module name and vendor string, tweaking the support message.
Dealing with NoAccel before Rotate, since the latter overrides the first.
Whitespace normalization.
Some small tweaks, to bring trunk and randr branch a bit more in line.
Whitespace and formatting adjustments after the libpciaccess merge.
Some comment and whitespace tweaks, in both trunk and randr branch.
Overwriting svnversion.h for a release, instead of using a separate file.
VM800/CN700 appears to have a CLE video engine (reported by Vincent Magnin).

Cyril Brulebois (8):
Add DM-Upload-Allowed: yes, Julien is doing a good job, let's allow him to perform uploads.
Upload to experimental.
Remove upstream files.
Add upstream files from 0.2.906 tarball.
Bump changelog.
Deal with X 1.12 by calling vgaHWSetStdFuncs().
Add xutils-dev build-dep.
Upload to unstable.

Dave Airlie (2):
openchrome: update to new API.
openchrome: fix issues with compat-api against latest server.

Gabriel Mansi (5):
Added missing includes. Check for null data block in vbe. Renamed VT3157 to VT3324
Hasee F700C (reported by Trent W. Buck)
Fix pause address register for CX700
Initial XVideo support for VX800
Initialize CRTC before a mode switch. Fix bug #260.

Harry de Boer (4):
Fix VT1625 output sensing. VT1625DACSenseI2C was using the same code as VT162xDACSenseI2C but the DAC sensing bit is in a different register for the VT1625. Also adds support for the VT1625S which has only four DACs.
VT1625 register count is 0x82
Set possible_crtcs to allow tv outputs to be connected to crtcs.
Return valid DisplayModePtr list for tv outputs. When creating a list of DisplayModePtr the ->next and ->prev pointers should point to the next/previous item in the list or only the first modeline will be recognised. This patch changes via_tv_get_modes to use the xf86ModesAdd and xf86DuplicateMode helpers to create correct modeline lists from a DisplayModeRec array.

Ivor Hewitt (46):
Initial import
Bring unichrome vt1622 mode changes over.
Split libxvmc into pro and non-pro
Pro split tidying
Oops missed a file
Installing made libraries would be smart
Fix library selection
Removed debug message
vim
Include John Wier's extended FIFO fix... (and set editor command correctly this time)
Additional card-ids. Patch from Xavier Bachelot
Added MSI K8MM-V id, reported by Philip Prindeville.
commit 90fd52afe1d7996c24cce996c7982e3948ec947c
Merge fixes from xorg for typos in mode setting and sync set
Initial steps to incorporating CH7xxx and VT1625 support
AsRockl device id addition
More device ids
Bug #4855: Fix VT3122 and VT3205 dotclock PLL setting. (Luc Verhaegen)
Mitac 8889 device id (Daniel Schindler)
Forgot changelog
Add vm800 bandwidth.
Dont kick vblank with dri disabled (Joris van Rantwijk).
Fix pro-mode and remove pro warning. (Paul Bender)
More device ids
Device id
Update email addresses
Just another id. (Tobias Blom)
Add svn version number onto driver message text.
Allow generic card-ids, although it should be possible to remove the id lookup completely.
Revert pointless change. :)
MaxSelect Optima deviceid - Anton Farygin
VM800 Xv support
Add RV32 too.
Minor code shuffling
Shut luc up.
Checking svn still works.
Unbreak IMakefile
Packard Bell imedia id - Jean-Damien Durand
Apply xv abi compatibility patch from morgoth
tweak abi check
Implement pro-clock-calc for non-pro
Oops
First auto-generated tv mode for vt1623
Attempt to fix NTSC noscale and tweak PAL mode too
Remove warning
Minor compilation fix

James Simmons (265):
Branch created to implement future support for TTM and KMS
move UMS code out of via_driver. It will make easier once KMS is implemented
Start of placing UMS code seperate from via_driver.c
remove RandR from UMS. This will be shared between both UMS and KMS. Expose VIAUnMemMap, the plan is to move that to UMS since the TTM version will be different.
move vga text mode restore and save to vgahw file.
Remove UMS specific code from VIAPreInit.
Test for when vgaHWPtr is missing to avoid segfault. This will be the case for KMS.
UMS DPMS code moved to crtc code.
cleanup of VIAScreenInit. It shoudl be usable for both KMS and UMS now.
remove some more cruft.
fix for VT switching.
needed to break up UMSAccel. Model more after the nouveau instead of the gallium xorg server.
Merrged to trunk at 925
prepare for handling of dri1 and dri2 apis
Seperate out DRI handling from UMS/KMS handling. The idea is to allow DRI2 support under KMS or UMS.
test to see if we use dri2 or dri1 api.
enable EXA by default
start sorting out ums form kms
missed a file
reorder code for crtc handling.
merge cursor code into crtc handling
ViaValidMode is UMS specific
revert changes since r930
reorg to move over to xf86CrtcFuncRec
abstract palette handling so it can work with both KMS and UMS
Move dpms to xf86Crtc
link in cursor into xf86Crtc
If we don't detect any drm driver then just run in pure slow UMS mode.
allow dri1 support for via kernel module version 3.
rename via_mode to via_outputs
Both SwitchMode and AdjustFrame are now wrappers about the crtc handling functions
replace home made timing with standard xf86 routines
cleanup white scapes. Fill in screen info for crtc and output structures.
place vgaHWSaveScreen in the DPMS handling routines. This allows us to replace VIASaveScreen with the generic xf86SaveScreen
move DPMS and SaveScreen to generic layer.
VIACloseScreen can call UMSLeaveVT since they preform the same functions.
Have CloseScreen call LeaveVT to simplify the function.
white space cleanup for outputs
Create analog output
Moving LVDS over to Ouput RR api.
Hanld ecase where no LVDS exist
Move DP hardware ot Ouput format
move TV over to xf86Output RandR.
Implement save and restore function
Move some accel handling code to the UMS section since this will change with TTM handling
UMSResource is alot more generic
incorporate VIAWriteMode into Crtc mode_set. Call xf86CrtcSetMode when we enter UMSEnterVT
set the output->crtc field to specific CRTCs
For some reason TMDS crtc mapping breaks my laptop.
move over to real mapping between CRTCs and Outputs
xf86CursorInfo is now a wrapper around the xf86Crtc cursor functions.
Create two different crtcs to handle each IGAs.
Move ouput mode setting outside of crtc mode handling
VBD dpms cleanuip for Outputs and Crtcs
Oops mixed up argb and mono cursor functions
Fixed the issue that was breaking my desktop environments. Mode setting more than juts rte CRTCs so calling mode_set directly was a bad idea in the ScreenInit function
Xaviers patch to turn off Legacy mode switcing in older hardware.
Use xf86_cursors_init. With this now the cursorstr code is just awrapper around the RandR crtc cursor functions.
A little more cleanup.
set desired mode for each Ouput
Old style TVs can't get their modes from EDID. Also return the true mode of the LVDS
Another screen function needed for RandR.
Remove extra cursro variables.
Placed ViaCRTCInfoPtr into the driver_private are of xf86CrtcPtr
Fix LVDS support.
Discovered problem is not crtc to output mapping but the ViaModeSecondCRTC code is broken for setting the mode on the second IGA. So instead we have to use the code for both CRTS to work the second IGA
Handle CRTC limit checking in crtc randr layer.
Remove ranrd hooks, instead use default behavior.
Moved the VIACRTCInfo into xf86CrtcPtr.
Properly handle vtSema and FirstInit.
Moving to have the VT enter and leaving code generic since we hide the UMS and KMS code behind rthe RandR api.
move vbeDoEDID to to UMS handling code. Change order of detection to avoid LVDS TV conflict.
Use Randr instead of our custom turning devices on and off.
THe CRT analog hack is not needed. VGA wil be detected if present
Made a more generic FreeScreen. Add a test for KMS support. Test DRM right away.
remove DGA since its been replaced by RandR shadow functions and shadowfb
shadowfb has been replaced by Randr shadow and shadow api.
Remove VIAInternalScreen. Merged into ScreenInit
Moved MapFB to UMSPreInit. Accel code white space cleanup. Don't call viaInitXAA if we are using EXA
Moved mapping and unmapping to close screen and screeninit. Some dri code in via_ums.c moved to via_dri.c
Cleanup between via_ums and via_accel so we can support GEM and non GEM memory management
dri protocol is not returned VIADRIFinishScreenInit
drmSetMaster and drmDropMaster should only be used with KMS
output cleanup in via_video and vgahw
proper debug info
Must test for DRI1 support since those methods don't work well with TTM.
TTM and DRI1 do not mix
bump version
move libxvmc into src directory
rename via.h to via_eng_regs since the defines are related to the video encoders. Also this syncs more to the other via xorg branches.
sync via_reg.h to via last work.
DVI OutputRec in ViaBIOS is not needed
We don't need a special scrnIndex
rename Outputs so RandR options can work with KMS and UMS
First part of the merger with VIA newer code.
Cleanup for mpeg encoder handling to make merged with VIA code easier.
integrate VIA via_regs.h header.
moving to VIA newer cursor definiations.
remove useless conditonal discovered by Xavier.
Newer libpciaccess broke are xorg driver. Tracked down the breakage and fixed it.
organize to create a buffer manager
Make the native resolution of the panel the preferred mode to set to.
rename via_memory for memory allocation handling
Fixes from Thomas Schlichter. Declaring VIAFreeRec as static is incorrect. Wrong bool value can be returned from VIAEnterVT. Call VIALeaveVT directly in VIACloseScreen, can oops otherwise. Move VIAUnmapMem to VIAFreeScreen.
Change viaIsOffscreen so we can use it for EXA PixmapIsOffscreen function. Move viaSetupCBuffer into UMSAccelInit since it's only used by EXA layer. Also by moving it viaFinishInitAccel will not go to a NoAccel state now. Final improvement is in UMSAccelSetup we don't clear the screen since we already do it in VIAEnterVT.
Makefile file to build via_reg_tool
Make AGP setup DRIInfo independent. Will wrap inside a memory manager.
split xaa and exa support.
Handle the case where the reported VRAM is greater than the maximum amount exposed by the pci bar
cleanup framebuffer handling for pre TTM code
Fix for xorg 11.2 ap change to vgahw
Make DRI buffers DRIInfo free for early allocation
Moved allocation to memory manager. Actually fixes cursor on VX900
add in support of I420 video format
Fix typo for cursor allocation
Tab and white space fixes.
Return bool for InitXAA. Also use CreateScreenResource
Move DRI1 and no DRI case to new buffer allocator. Some more work needs to be done to get it ready.
Sync the version across all files by the magic of autoconfig
Implement drm_bo_map and drm_bo_unmap. Have hardware cursor use it.
Always pass ScrenInfoPtr to every buffer management function. Cleanup buffer_object alot plus use the handle for non accel case.
Move DRM master node opening and closing to the ScreenInit and ScreenClose
Move shadow buffer handling to CreateScreenResources
More buffer management cleanup
Cursor initalization is DRM api type independent.
free the cursor when closing the screen
Cleanup dri layer to use buffer manager
Add in which domain to allocate memory from
No longer used by exa layer
Move exa scratch buffers to drm buffer manager
Compile fix that was missed.
Make exa allocation even generic
better debug plug fix not freeing drm memory
Set the buffer virtual pointer.
Convert video handling to buffer manager api
Test if scratch buffer exist. Fixed typo
Another missed NULL check for the texture buffer
bug fixes from Thmoas for the allocator
remove dual backends. Just use the FBManager for the non-accel case
merge panel and lvds code into one spot
Cleanup VIAFreeRec and a casting issue
actually remove via_panel.c file
Always give only have of the vram space to the DRI1 layer for allocation. The other half can be used for EXA.
More debug info for memory allocation
Handle agpDMA flag in DRI code.
Move display surfacet to buffer manager.
sync with drm-openchrome and libdrm
add in aligment and map handles to the buffer manager
Initial support for memory allocation from TTM/GEM layer. Not ready for prime time yet
fix bug for ticket 422
name via_bios.h to via_ums.h
1366 mode is not a valid CVT mode so we use the closes one which is 1368.
fix wrong offset to registers for HQV engine
fix a double free and unmapping of video plane memory
Support svn and git version handling
ignore specific files for git
ignore man files
More files to ignore for checkin
Support versioning bug git or svn
Turn ViaSimultaneousInfoPtr into a simple Bool. Its not even used.
Make ViaPanelInfo the private data for xf86OuputPtr
Sort out the LVDS code. Now we have lvds_detect get the EDID or the native mode index. For the get_mode function the DisplayMode is generated from the EDID or the NativeModeIndex. The Legacy code also has start to blend together with the standard mode setting code.
Makefile file to build via_reg_tool
The output prepare functions are done at mode setting time. With RandR you can setup up various combos of outputs and crtcs. This is done before we attempt to set a mode so when the mode actually does get changed we programming the mapping
Fix detecting the LVDS more than once. The status was never set to connect because we only handle the case when the PANEL_INDEX is valid. Also it appears LVDS-0 are always mapped to IGA2 except for the VX900.
fix that prevented RandR mode switching
Modes are validated by the RandR layer so no need to call xf86ValidateMode.
Fixes for non accelerated systems with the recent changes for the mode validation cleanup.
Merge branch 'master' into kms_branch
passed wrong flags to gem allocation ioctl
Sync to drm-openchrome for domain and permission change to api
KMS mode setting support
Update alloc flag form memcpy test
drm_bo_alloc allocates a raw chunk of memory. A new function drm_bo_alloc_surface had been created to allocate different types of drawing surfaces.
Currently register saving is done in both the ScreenInit and EnterVT function. By the time we enter the EnterVT function all vga text register setting are lost since the origin text mode settings are over ridden that where gathered in ScreenInit. Now its all done in EnterVT and the EnterVT function is called by the ScreenInit function.
Needed for non acceleration mode
Probe the hardware for the resolution of the LVDS instead of depending on the scratch registeres
remove accel blanking screen. It was being called even with NoAccel for some reason with drm-openchrome kernel
The presence of KMS is not a error so reported as info
Fix for build on embedded platforms. From bug 430
Fix to building with dri.
Allocated wrong data structure
More files to ignore for git
More iga mapping fixes. Some old functions from th pre randr days where over riding the new iga code.
The patch for building on embedded platforms sets the XF86DRM variable for you so this exposed a bg where we were calling VIAInitalize3DEngine in the EXA startup code but that code was in via_dri.c which was not compliled in for the VX900 case. Move VIAInitalize3DEngine to via_ums since its not needed for the KMS/TTM case.
Merge in VX900 video support from trunk
More fixes for DRI. Also add support for UDEV for hotplug events.
Fix from Schlichter so xorg driver can be built on embedded platforms
Using xorg conf options with the panel casues a segfault. This fixes the segfault by allocation the necessary memory to store the data. Also fixed some memory leaks.
Filter out I2C bus. This is needed for the XO which hangs when you probe the second I2C bus.
Removed FBBae dependency of the EXA layer. The frame buffer could change due to a resize.
The framebuffer address might of changed before we do a mode set so we must tell the display where the new drawing surface is at.
Merge branch 'kms_branch' of git.freedesktop.org:/git/openchrome/xf86-video-openchrome into kms_branch
Fixed the cursor for KMS to actually display
Disable EXA composite since it is broken. Will fix at a latter date.
Fixes to allow mode resolution change on KMS
Broke unaccelerated mode memory allocation with last commit
Moved to EXA only
Fix object memory freeing for all cases
Don't do a memset on a NULL pointer for the ExaDriverPtr allocating
drm.offset from api one gives the absolute address instead of the relative address to the vram start. Also fix the default source for non acclerated mode
Fixed the EDID handling for the LVDS. The code was to dependent on a NativeMode being present which is not set with EDID.
Fixed the crash that happened when you logged out of the session. It wass both initialize and restoring the video state on the call to EnterVT which locked up the system. This fix also restored the DMA command buffer
Borrow the ROP values from xf86-video-chrome which comes from the nouveau project. Its silly that every one have their own ROP values but its the way it is. This allows the complete removal of XAA.
Borrow the ROP values from xf86-video-chrome which comes from the nouveau project. Its silly that every one have their own ROP values but its the way it is.
Death to the xaa layer completely
enable EXA by default
Disable EXA composite since it is broken. Will fix at a latter date.
Fixed screen resizing from UMS and KMS.
OLPC needs a very special mode so we hard code it. Also we provide the physical size of the display since the OLPC wants it.
Handle mmap failure case.
Pass in the pitch variable for allocating a surface. Will be used for EXA dynamic allocation
Rename video and swov to xv naming.
Pass in the crtc to know what to attach the video plane too
Hardware cursor does work now
Patch from Bartosz to fix I420 on CME engine.
Patch from Bartosz to fix I420 on CME engine.
Remove all traces of mono cursor. The xf86cursor layer has a mono to arbg mapper we can use by default.
Remove no longer supported xf86PciInfo.h
EXA has severe memory leakage.
Separate out UMS accel from exa for when we support EXA with KMS
Remove obsolete code
Move TwodRegs to ViaCommandBuffer
Untangle the EXA layer. Moved the UMS specific code via_ums. Started to lay foundation for H5 and H6 support.
Currently the xorg driver supports only H2 dma packets but newer hardware doesn't directly so implement H2 with a different hardware set. This patch separates the H6 hack and H2 layer to allow ease of implementing real H6 support.
Unwind some of the XAA helpers into the respected EXA functions
Sync names to KMS names so config files don't need to be changes in the future
Xavier patch to add the git branch to the generated version.h file
The EXA corruptions people are reporting appear to go away when you set the MigrationHeuristic to greedy which means all pixmaps have their buffers pushed to the frame buffer. We force enable it as a work around. Some day it will go away
ViaDoubleCheckCLE266Revision needs to be called later when the register stack is fully setup. Also this function should be called for UMS and KMS case were before it would only be called for UMS case.
Remove debug message and some more white space cleanup
Compile cleanup with stricter gcc settings
KMS could not mmap on 32 bit machines due to off_t issues which was resolved with the AC_SYS_LARGEFILE macro. Separated out the via_drmclient.h header from the via_drm.h header. Synced to drm-openchrome kernel changes for gem creation.
Minor fixes that makes it compile with -Werror on x86_64 from Natanael Copa
Force EXA to always be greedy
Patch from Xavier to inform people to where to submit bugs
Merge branch 'master' into kms_branch
Since DP handling is not very functional disable it unless forced by ActiveDevice
Merge branch 'kms_branch'
Fixes for make distcheck to work
SCRN_INFO_PTR no longer works in new xorg servers when used from
When setting a new video mode we were assuming the offset into the frame buffer was 0,0 which is not always the case
The default timeouts for values for I2C are way to small for the defaults xorg server sets. We set them to the VESA DDC specs values instead.
Setup the basic IGA registers. This is needed to allow CLE266 to work in Non Legacy Modesetting.
ViaDisplayInit called at startup broke VT switching. Instead I call ViaDisplayInit in VIARestore which allows CLE266 to work without legacy mode
XSERVER_LIBPCIACCESS is now decrepide but we need to handle older platforms (rhel5) so HAVE_PCIACCESS was created. This patch allows us to use libpciaccess 0.11 or better or work around the lack of libpciaccess.
Sometimes we fail to get a EDID from a VGA monitor but a real display is attached. In this case we look to see if a vblank is raised by the analog monitor and report it as detected
Do not support earlier libdrm versions that lack KMS support.
Fix a double free when exiting the xorg server
Doh, shouldn't code when sleepy. Free the TV i2c struct only we still have a pBIOSInfo struct left
Fix a long standing issues of detecting the proper amount of VRAM on the P4M800 devices. We can tell it is a p4m800 versus a km400 by what pci bridge it has and reading the proper pic function
As the comment suggested in via_outputs.c probing the 3rd i2c bug locks up the P4M900. So we re-enable the via_card_id handling of what outputs to test for to get around this problem.
Add Nec Powermate VL5 ViaCardId data to support this device
Support older X servers for RHEL5
Allow via_output.c to build without C99 mode
miInitializeBackingStore no longer exist in xorg server 1.14. Removing causes no regression in earlier versions from my testing
Rename GEM data structure in the fashion of other drm driver apis
via_xf86crtc_resize can handle both UMS and KMS modes so we can remove drmmode_xf86crtc_resize.
Enable xorg driver to recieve hotplug events.
drmmode_set_mode_major does this for us so no need to do it here
Integrate the different formats now supported by libdrm. This makes allocating buffers for KMS planes in the future much easier
In our driver in order to support both KMS and UMS we can't call drmmode_set_mode_major directly but instead
Fix the problems with UMS multiple screen handling. KMS still has problems
Patch from Krummenacher <max.krummenacher@toradex.com>
Our xorg driver has moved to using drm_fourcc.h but not all distros have a new enough libdrm so building can fail. We place a copy in our source tree to ensure building and this header doesn't change to often

Jon Nettleton (34):
Add automake files to allow running make from the root directory
remove binaries
Undoing change committed in r212. Sorry
fix for compiling with mesa 6.5.1
fix some config options that are negatives
remove different includes based on xorg version check
forgot to change Changelog date
migrated per chipset default options from experimental to trunk
removed to chipset definitions that haven't been merged back to trunk yet
Initial VX800 support thanks to work mostly done by Xavier and Harald
finish acceleration changes for the VX800 chipsets
fixed VX800 so ActiveDevice is no longer needed
Added more needed modes to ViaPanelModes
Fixed Crash worked around by XaaNoImageWriteRect
Clean up compile warnings
fix placement of pVia->FrameBufferBase to MapFB
Merge work done in hwcursor_branch back to trunk
adjust monitor values to handle preset LCD panel and TV modes
Temporary workaround for freedesktop bug 21563
fixed viaFlushPCI wait for engine idle in vx800
redo removing the check for 3D_ENG_BUSY when flushing the pci bus and waiting for idle
fix 2d initialization for P4M900
remove EXA ifdef logic - we are too old for that
minor bug fixes and tweaks
small bug fixes for XAA and EXA
XO-1.5 panel patch and dot clock fixup. Patch by Xavier Bachelot
Remove incorrect entry from the Dot Clock table
XVideo support for VX855
Add timeouts to HQV while loops
Improve 2d performance on chipsets without AGP/PCIe
Add option I2CDevices
Disable Hardware Clipping for the VX855
Re-Enable the VQ for VX800/VX855
Fix XV crash on PM800 post VX855 rework

Julien Viard de Galbert (1):
Fixing git build package config

Maarten Lankhorst (11):
Revert "Add upstream files from 0.2.906 tarball."
Merge branch 'release_0_2_9' into debian-unstable
Merge branch 'nasty-unstable' into debian-unstable
Move to git.
Merge branch 'debian-unstable' into debian-experimental
Merge branch 'upstream-experimental' into debian-experimental
Bump changelog
New upstream release.
Merge branch 'debian-experimental' into debian-unstable
Merge tag 'release_0_3_3' into debian-unstable
bump changelogs

Pascal Ermster (1):
Change AM_CONFIG_HEADER to AC_CONFIG_HEADERS

Thomas Hellström (116):
* Brought in a stripped version of the client side of the DRI protocol.
* Fixed an undersized BEGIN_HEADER6_DATA command
* Fixed up X server XvMC with adaptors for different chipsets.
* Fixed missed overlay update in XvMCPutSurface when new source or dest rectangles were used.
- Added ChangeLog to the repository.
- Added the prepare-ChangeLogSVN.pl script for generation of ChangeLog entries.
Added better AGP control to the DDX.
Changed DRI client function names since they clash with via_dri.so,
- Updated incorrect field order in libViaXvMCPro.
Fix TV-out on Unichrome Pro. To make it work CRT+TV needs to be
Update via_dri.c to work with new xorg libdri.
Fix up the vt1623 TV encoder:
Fix XvMC on 64-bit systems.
- Free dotclock on Unichrome Pro. Pls report problems with
-Clean up the VIDEO_DMA ifdef mess before xorg inclusion.
-Avoid calling DPMS to blank screen during init with VBEModes
Fix Xv StopVideo for all ports and adaptors.
Fix Xv panning.
Fix segfault on EnterVT when DRI is not enabled. (Joris van Rantwijk)
Fix Xv panning also for VBEModes
Destroy video surfaces on video exit / stop.
Big Xv update.
Fix HW Cursor memory allocation.
Make via_video.c compile without -DXF86DRI. Broken with the
Sync dri handling with Xorg. This should hopefully make it run
Fix unichrome hangs with unsupported dotclocks (Luc Verhaegen)
Fix server crash when compiled without -DXvExtension.
Enable DMA Xv image transfers also for PM8x0 / CN400.
Revert the hardware cursor fix since it for some strange reason makes hw
Fix stray include left over by previous commit.
Second attempt at fixing the hw cursor issue.
-Fix compilation errors caused by the drm / dri update.
- Merge exa_branch from revision 52 to 67. For a detailed Changelog,
- Merged changes in exa_branch revisions 67 to 74. For a detailed
Updated ChangeLog.
- Fix Xv surface destruction and re-allocation.
Merged exa_branch changes between revisions 74 and 84. Mainly indentations and minor bugfixes.
Merged exa_branch between revisions 84 and 86.
Include the correct drm.h file. This is important for 64 bit builds.
Yet another drm.h include fixup. For 6.8 and earlier?
Merge exa_branch revisions 86:91
Add drmCommandWriteRead to drm loader symbols.
Remove drmCommandWriteRead from drm loader symbols again.
Merge exa_branch revisions 91:104:
64 bit fixes on libviaXvMC*
Fix trac item #13.
Adjust for modular build.
Removed Makefile.in since it is created by the build process.
Moved memset to blank screen from ModeSetting to ScreenInit, in
- Moved the memset from previous commit to after dri is initialized, so
- Protect the accelerated initial sceen clearing with a DRI lock.
Make sure the accel marker system is properly initialized.
- Modular build two identical XvMC libs. The pro lib was not a pro lib.
- Fix Xv YUY2/RVXX dmaBlit stride.
- Update Xv blit to new blit-combining feature of DRM.
- Merge in exa_branch from revision 104 to 127. Exa Composite acceleration.
- Modular build fix (Boris Dusec)
- Merge exa_branch 127:130. Rendering errors.
- Merge exa_branch 130:132. Rendering errors.
- Changelog got left over.
- Mark already destroyed XvMC contexts as bad, in case clients
- Fix planar Xv Flipping and surface allocation size.
- Exa branch revision 132:138 merge.
- Add a framebuffer map for DRI. Don't just assume that the
- run indent on libXvMC. Sync XvMC with Xorg 7.0.
Fix HQV offset bug.
- Added vgaHWGetIndex to vgaHW symbol list.
- Added a "NoXVDMA" option to disable PCI DMA for Xv image transfers.
- Make the Xv viaWaitVideoCommandFire loop time out. Better than server hang?
- The previous commit causes problems with older X versions.
- Import Luc's fix for sometimes-blanking CRTs on CLE266.
- Merge exa_branch revisions 138:152
- Removed time.h and sys/time.h includes from via_swov.c
- Merge EXA branch revisions 152:155
- Fix VBE refresh setting.
* libxvmc/Imakefile:
- Possible fix for XAA bug rendering solid lines as dashed.
- Support EXA 2.0
- Update compatibility check for XvMC.
- Update the built-against macro to recognize Xorg if present.
- Backport Makefile changes form xorg.
-Fix severe dri memory allocation bug. (Reported by Trevor Kramer).
Fix dri close screen segfault caused by previous commit.
Increase video command fire timeout.
Update DRM version checking.
Remove the "config.h" includes from libxvmc, as they break it.
Change DRM major compatibility from 3.x.x to 2.x.x.
-Changes donated by "PConRails, LLC" and "Tungsten Graphics, Inc.":
-Changes donated by "PConRails, LLC" and "Tungsten Graphics, Inc.":
-Fix false AGP detection and make sure we have the correct AGP
...And get default AGP size correct as well.
- Re-enable Ivor's Dotclock computation.
Fix a mutex unlocking issue (Luc Verhaegen).
Don't include drm headers in non dri build.
Removed some stray debug messages.
Remove VT3314 (CN700, VM800 etc.) mpeg2 AGP DMA support.
Via managed to mix up the 2D / 3D engine busy status bits.
Remove unused options.
Use a more aggressive fifo setting for CLE266 Cx.
Fix the dma2d option setting.
With EXA + 3D we need to reinitialize the 3D engine after
Back up DRI offscreen memory before leaving VT and restore it
Fix up previous commit for > 16MB sizes.
Force 3D context upload after a VT switch.
Now that we initialize the 3D engine every time we write
Proper register names in VIAInitialize3DEngine.
Cause a "FatalError" if EXA is used with an un-patched Xserver,
Reimplement download from screen to something more easily
Fix download from screen which was reusing kernel argument
Thomas Hellstrom <thomas-at-tungstengraphics-dot-com>
Blank and clear screen when entering VT.
Initialize the AGP ring buffer before command submission.
Use non-power-of-two aligned textures when possible.
Fix AGP mode setting.
Work around an obscure hardware limitation when texturing from
Fix possible line clipping bug.

Timo Aaltonen (5):
Merge branch 'upstream-unstable' into debian-unstable
bump the version
Remove 10-Deal-with-xserver-1.12.diff, not needed anymore.
control: Bump policy to 3.9.3, no changes.
close a bug

Xavier Bachelot (247):
More pci-ids and some fixes
Replaced mainboards with borrowed pci-ids by generic chipset name
Added Averatec 3715
Sort K8M800 pci ids
Asustek P5V800-MX
MSI PM8PM
Shuttle FX21
Two new pci ids
Gigabyte GA-8VM800M-775
Gateway MX3210
little configure.ac fixes
Packard Bell Easynote R1100 (from unichrome.sf.net)
MSI PM8M-V and MSI PM8M3-V
man page fixes
ECS P4M800PRO-M (Amy Overmyer)
Mesh Pegasus (Peter Lord)
2 more pci-ids
fix stupid typo :-(
2 more pci ids
Fujitsu/Siemens Amilo L7310 (Andris Jankevics)
Fujitsu-Siemens Amilo K7610 (Botos Pavel)
Epia EN (Olivier Sessink)
Biostar M7VIZ (reported by Daniel Robbins to unichrome)
DDR533 support. Fujitsu/Siemens Amilo L7320 pci id.
Fujitsu/Siemens Amilo Pro V2055 (reported by Teifion Siwft)
another VM800 pci id (reported by jdoe)
manpage makefile fix (Michael Moerz)
RoverBook Partner W500 (reported by Rikz)
Clevo M5xxS (Reported by Thomas P Crawford)
ECS C7VCM (reported by Mark Huijgen)
720x480[Under|Fit|Over] modes are actually 480P, not NTSC
add 720x576Over mode for VT1625
VT1625 480P fixes (Tim Dodge) and table indentation clean up
fix VM800, CN700 and P4M800Pro all reported as VM800. They are now reported as VM800/CN700/P4M800Pro.
Clevo/RoverBook Voyager V511L
Clevo T200V (David Numan)
fix typo in last commit
Asustek P5VDC-MX (reported by Alberto Salmerón Moreno to Unichrome)
fix typo in vt1625 modes table (trac ticket #105)
real Asustek P5VDC-MX and fix Asustek P5VDC-TVM
Backport 2 fixes from Xorg tree
2 new ids
one of the 2 pci ids is a CLE266 and thus needs to be in trunk
Biostar ARKplus III pci id
rebase to release_0_3_0 branch rev. 413
Add VT1625 720x576 PAL mode (Daniel Ramsay)
Add VT1625 720x480Under, 720x480Fit and 720x480Over NTSC modes (Ken Huisman)
fix an error in VT1625 720x480Under NTSC mode
Improved P4M890 memory detection
Indentation fix
Add a COPYING file and a license header to via_priv.h
revert changeset 401, test is not valid
one more P4M900 board (reported by Marco Pieruccetti)
fix copyright header
report chipset too in case of an unknow card id
K8M890 can use DDR667
FIC CE260 pci id (reported by Damian Kolkowski)
Panel on CX700 needs VBE
PCChips V21G pci id (reported by Edward Allcutt)
New pci id for Biostar P4M900M-M7 SE
Bump driver version
Bump for real this time...
MSI Fuzzy CX700/CX700D (Reported by Forest Bond)
Fix missing newlines
Packard Bell EasyNote XS (aka Everex Cloudbook CE1200V)
ASRock P4VM900-SATA2
Merge P4M900 fifo setting fixes from randr branch
Asustek P5VD2-VM SE (reported by J. Dale Gonzalez)
Biostar P4M800 Pro-M7
Medion Notebook MD96483
VIA Epia SN (reported by Stefan Bellon)
Apollo BMOVE SR (probably a rebadged Clevo laptop)
Clevo M660SR
Fix Xv on LCD for VM800
Mitac 8515
libpciaccess support
Mitac 8624 aka Benq Joybook R42 (reported by Stepan Zatupov)
merge changeset 436 and a fix from randr
yet another tiny sync with randr
Create src/svnrelease.h on make dist
fix bug report link and remove a superfluous newline from the generated svnrelease.h
Disable AGPDMA for VM800 by default
Prepare for next release...
Asustek M2V-MX SE
Bump version for release
include svnrelease.h in tarball
hopefully fix svnversion.h once and for all...
print driver version in the libpciaccess path
fix chipset revision detection in libpciaccess code path
FIC CE261 (reported by Robert Henney)
Fix Xv on LCD for CX700
some new pci ids
replace xf86strstr by the unwrapped version
new pci id: Mercury P4VM800M7
Disable XvDMA for P4M890 and K8M890, it is broken...
Samsung Q1B (reported by Massimo Vighetti)
fix Xv on LCD for K8M890
put an end to the Xv on panel issues
fix typo
re-enable XV on panel for VM800
fix interpolation for CN400
MSI VR321 (reported by Kot Cheshirsky)
ECS CLE266 (reported by Mohammed Abd El Hameed)
Foxconn P4M9007MB-8RS2H (reported by Stephen Yu)
Simplify memory bandwidth setting
0.2.903
MSI K9MM-V
update NEWS with the release date
Gigabyte M704 / RoverPC A700GQ
remove support for pre-xorg 7.1rc1 Xv ABI
IBM AnyPlace Kiosk 3xx (probably a VIA board, but I can't find a better reference)
Add missing 640x480 PAL mode for VT1625
merge Gabriel Mansi's changes from randr branch plus a few additionnal not commited fixes
oops, forgot to svn add the new files :-(
disable ARGB cursor for K8M890
Allow FullHD with DDR400
VIA VT8454B has a panel
MSI K8M Neo-V (broken pci id, there is already one with a good pci id)
minor typo
redo a couple changes lost in initial VX800 support commit
another Mitac 8624, but with a P4M890 (possibly 8624P or the other with the P4M900 is 8624P ?)
VIA Epia M700
Hide overlay when video is invisible (ticket #266)
minor typo fix
Axper XP-M8VM800
properly link XvMC (patch from Julien Cristau)
XvMC symbol visibility (patch from Paulo Cesar Pereira de Andrade)
XvMC Unichrome Pro allocation fixes (patch from Daniel Drake)
restore panel backlight after VT switch (Ticket #29)
Disable ARGB cursor, it's not working properly yet
Twinhead M6 (reported by Bradon Kanyid)
set P4M890 primary FIFO (from ttm branch, changeset 733)
Samsung NC20 (reported by Matthieu Herrb)
ViaPanelGetIndex: 1 bugfix and 1 cosmetic fix
Fix 2D engine init (patch from Bartosz Kosiorek)
CX700 integrated TMDS (Patch courtesy Tim Chen from VIA).
Foxconn P4M800P7MB-RS2H (reported by Onno Steenbergen)
VX800 integrated TMDS support
don't try to enable XvMC on VX800
Fix use of uninitialized variable (patch from Bartosz Kosiorek)
Fix possible use of uninitialized variable (ticket#292)
Fix DFP pad mask
Cosmetic fix for ViaMMIODisable (patch from Bartosz Kosiorek)
MSI P4M900M3-L (reported by Jonathan Ernst)
make sure Chrome9 chipsets use software rasterizer for 3D
remove VT8454B LCD flag, it breaks stuff and I'm unsure about the board anyway...
clean up duplicate defines in via_regs.h
fix null pointer dereference in viaExaCheckComposite (patch from Daniel Drake)
fix a segfault on shutdown in ViaCleanupXVMC when there's no Xv (patch from Daniel Drake)
fix cursor on secondary (patch from Bartosz Kosiorek)
Sharp PC-AE30J
VX855 basic support (patch from Harald Welte)
Modify PLL generation and add VX855 support (patch from Chris Ball and Mitch Bradley)
Switch on LVDS pads only for active channels (patch from Peter Stuge)
make sure VX855 uses swrast
fix dotclock table ordering
remove loader symbol lists
Fix hardlock on resolution change (patch from Rafał Miłecki)
disable TMDS by default for now, output detection is not working properly
Build fix for xextproto >= 7.0.99.1 (Ticket #312)
Haier A60-440256080BD (actually an ECS mobo, ticket #311)
Lenovo S12 (ticket #316)
fix up detected hardware reporting
xorgconfig is gone...
ECS P4M890T-M v2.0
revert changeset 764
update for resources/RAC API removal
fix too long lines in man page
Packard Bell Lima (ASUS MBP5VDZ-NVM)
properly enable MMIO and pci burst on all HW. Fixes #323
update NEWS to prepare next release
better check for resources/RAC API
add missing comma (cosmetic)
ModeSwitchMethod option (patch from Rafał Miłecki)
fix for #326 (hopefully...)
last doc tweaks before release
bump release
Guillemot-Hercules ECafe EC900B (reported by Andrew Jones)
OLPC XO 1.5
Twinhead K15V (from ticket #347)
FIC PTM800Pro LF (ticket #351)
MSI K8M890M2-V (ticket #356)
fix DFP parameter description for the ActiveDevice option
Disable AGP and DMA by default for VX800 and VX855
Revert 2 changes that should not have gone into rev 846
ECS P4M800PRO-M2 (V2.0) - reported by Jandré le Roux
fix typo in r857
fix a typo and add a FIXME following the VX900 merge
add VX900 to man page
FIC CE2A1
MSI PM8M-V
Lenovo ThinkCenter E51 8714 (reported by José Jorge)
Sharp Mebius PC-CS30H
MSI PM9M-V (reported by Florin Rentea)
revert CLE266 to legacy modeswitch
reset patchlevel to 0
Hewlett Packard DX2020
Import Harald Welte's registers dumper code (http://svn.gnumonks.org/trunk/via-chrome-tool/registers.c)
Add configure option to build the regs dumper
Add basic usage function for the regs dumper
fix build option
ASRock PV530
Handle X server 1.12 (videoabi 12) (patch from Adam Jackson)
Add support for I420 Xv surface
Limit video RAM size to 256 MB, PCI BAR can't handle more
Import Harald Welte's registers dumper code (http://svn.gnumonks.org/trunk/via-chrome-tool/registers.c)
Add configure option to build the regs dumper
Add basic usage function for the regs dumper
fix build option
prepare for 0.2.905
bump version to 0.2.905
fix typo
Don't load the XAA module when using EXA or NoAccel.
Prepare for 0.2.906
Fujitsu Futro S100 (reported by Ferencsik Péter)
update NEWS file for upcoming 0.3
HP Compaq dx2040 (reported by Tom Li)
HP Compaq dx2040 (reported by Tom Li)
fix 0.3.0 release date
Fix typo in bug tracker link
Install registers dumper tool in sbin rather than bin
Use bits description for Software Reset Control register
Remove a couple unused variables
Tweak release date again
Simmtronics SIMM-PC VX900i (reported by Sagar Sawant)
Lower libdrm requirement to 2.0
Remove --(enable|disable)-dri configure option, it depends on whether the X server was built with or w/o DRI support.
Dell Optiplex FX130 (reported by Maciej Kazulak)
bump version to 0.3.1
Fix typo in registers description
Biostar Viotech 3200+
Fix typo
Change to new mailing list address
Fix indentation
Fix registers description
Fix CN400 Xv on secondary
typo
When the card id is unknown, don't check CLE266 revision if the chipset is not CLE266
Fujitsu Futro A300 (reported by Alberto Asuero)
Revert CLE266 to legacy modeswitch
Add an explicit message for KMS support configure check
Hardcode panel size for the XO
bump version to 0.3.2
Biostar Viotech 3100+ (reported by J. Scott Heppler)
make iga[12]_crtc_mode_fixup more verbose when the mode is rejected.
Probe TV encoder on all but the P4M900, unless we know the board has a TV encoder.
Bump version to 0.3.3

Reply to:

Prev by Date: xserver-xorg-video-openchrome: Changes to 'debian-unstable'
Next by Date: xserver-xorg-video-openchrome: Changes to 'debian-unstable'
Previous by thread: xserver-xorg-video-openchrome: Changes to 'debian-unstable'
Next by thread: xserver-xorg-video-openchrome: Changes to 'refs/tags/xserver-xorg-video-openchrome-1%0.3.3-1'
Index(es):
- Date
- Thread