|
Issues
Resolved in Build 1187
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
BugID: 2293093,
2294228
|
Add "UseOriginalPitch" in CMD_SURFACE_STATE.
|
GMM
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® G965 Express Chipset,
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Support width
in DWORD for HW limitation.
Fix Description: Add new
member "UseOriginalPitch" to The CMD_SURFACE_STATE to be used
when surface is used in Media Kernel.
|
|
BugID: 2291779,
2287747, 2288939, 2325298
|
External XGA LCD does not turn on when LCD/CRT EDT ->
S3/Res.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: The issue is
happening to due to an unexpected call sequence from OS, where during S3/S4
events it commits the secondary first and then extends it to primary. This
code in miniport that derives the next topology (by taking care of
preserved paths) gets confused in this case and adjusts this topology in
the reverse order (LFP + CRT would now become CRT + LFP). On subsequent
commits the topology would become normal, which would now confuse GRM,
because the current primary and secondary have changed without proper
release notifications. Also, the code in miniport to identify valid plane
swaps would get confused because of the primary/secondary swaps. This would
result in valid plane swap recommendations to be ignored. All of these,
would result in wrong pipe/plane being programmed and there by results in
display blankouts.
Fix Description: The fix
provided is to make sure that the OS provided importance ordinals are
respected where ever possible, and to avoid checks based on importance
ordinals while checking for plane swaps.
|
|
BugID: 2222790,
2289839
|
Freeing cache space at FlushCommandBuffer time.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: BindPipeline
was failing on the first call after a flush because one of the state caches
was full. Because this was immediately following a flush, the clients were
not calling FlushCommandBuffer. This caused our copy flags, which inform
the KMD of data to be copied from the Command Buffer to the DMA buffer, to
become incorrectly set.
Fix Description: The fix
checks the caches at FlushCommandBuffer time to see if they are full, and
if so, it will free space then the next BindPipeline call will succeed. A
similar failure could potentially occur on Intel® G965 Express Chipset once
the vertex shader is enabled and so, we have made a similar fix for Intel®
G965 Express Chipset. There should be no game hangs that have been caused
by this issue.
|
|
BugID: 2222232,
2222651
|
DVD playback content flashing when HWA is enabled.
|
DVD
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® G965 Express Chipset,
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Add floating
point save/restore in DXVA.
|
|
BugID: 2221989,
2223076
|
Driver does not seem to support HDMI - No Video Code
avaiable on HDMI, incorrect AVI InfoFrame, and no 576p support.
|
SoftBIOS
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: VIC buffer was being set to NULL during boot up.
Fix Description: Removed BASEDISPLAY_FreeVICBuffer() function call from
BASEDISPLAY_FreeEdidBuffer(). Instead calling it just before retriving VIC
list and inside destructor.
|
|
BugID: 2221927,
2222802, 2254287, 2254668
|
Backlight hot fixes and fixing policy data structure
misalignment between builds issue.
|
Power Conservation
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset, Intel® GM965 Express
Chipset
|
|
Resolution Description:
Root Cause: OS switches
from Dual Display to LFP only in S3 transition first, then invoke
SetPowerState to S3. On Resume OS
resumed to Dual Display. As a result
the power conservation saved state is non-legacy PWM but resume to legacy
PWM. The Power Conservation policy structures got mis-aligned across driver
versions as features were added to the component.
Fix Description: Backlight
control is now completely taken over by driver in driver lifetime. Put in
checks for invalid data. Set policy data to default if invalid.
|
|
BugID: 2200000,
2222232, 2255368, 2256999
|
Add support for Post-processing Kernels optimized for
Intel® GM965 Express Chipset.
|
DVD
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Added driver
support for a new set of post-processing kernels implemented by the media
kernel group, based on the latest optimizations techniques.
Fix Description: 1) BOB kernels are no longer used. The
BOB functionality is now implemented using the scaling kernels and by
setting up the Vertical Line Stride (Surface State).
2) The ProcAmp filter is now placed after the Blending operation, before
Color Space Conversion; kernels where ProcAmp appears before scaling and/or
blending are no longer used. 3) Color Space Conversion is no longer a
standalone kernel, it is now combined with other kernels.
|
|
BugID: 2199988,
2199989, 2200000
|
PS3_0 Support for Windows Vista*.
|
Direct3D*
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Restrictions
on enabling pixel shader 3.0.
Fix Description: Removed the
restrictions on enabling pixel shader 3.0. This turned on the pixel shader
3.0 capabilities allowing pshader3.exe (The test used by dtm for Pixel
Shader 3.0). Also fixed an issue with PS3.0 Pixel Shader Constants where
the Masks were being generated improperly and implemented two missing API
callbacks For SetPixelShaderConstantI and SetPixelShaderConstantB which
were both not implemented.
|
|
BugID: 2197276,
2197240, 2197276
|
Restoring External Only results in black screen on both
displays.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: In Windows
Vista*, Miniport does not call SoftBIOS to set Primary Pipe for VBIOS [Bit
10 of 0x71414]. Therefore when a user boots with LFP only, VBIOS sets this
bit to indicate Pipe B [a value of 1]. On every mode setcall, we call
SetDisplayDevice with current pipe and plane and in the process SMI 050D
gets called with the associated Port-Pipe Matrix. The Pipe Index for this
is read from Bit 10 of 0x71414. Since this is set wrongly, the
corresponding Bits indicating CRT does not get set for the SMI. BIOS has an
embedded controller that will block the display from showing contents until
this SMI uses right values. Hence, display could not be seen on CRT only.
They do not have a similar logic for Int LVDS orDVI.
Fix Description: Call SB
before SetDisplayDevice so that the SMI has right values for Port-Pipe
matrix.
|
|
BugID: 2195831,
2199743
|
The display color turns to Default once aspect ratio is
applied.
|
SoftBIOS
|
Windows* XP, Windows Vista*,
Windows* 2000
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: PC in Windows* XP and SB gammahandler in Windows
Vista* enables gamma for non 8 bpp modes after programming the pallete
registers. For 8-bpp modes by default, it has to be enabled hence setting
the bit in plane enable only for 8 bpp modes.
Fix Description: Enable gamma in plane enable function only for 8 bpp
modes.
|
|
BugID: 2173293,
2197696, 2199616
|
Hot/warm dock, display switch has no function with CRT
attached.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: OS does not supply dock or undock status in its
docking notification. Due to this, we did not get the right dock status
from OS on dock. We dishonored the whole notification and did not update
internal driver status. Because of this all the subsequent ACPI hot keys
were considered as docking events, so we did not take any action on them.
Fix Description: Disregard
the OS notification altogether and use scratch pad mechanism to identify
dock and undock in the same style as in Windows* XP. Also added the
DXGK_ACPI_POLL_DISPLAY_CHILDREN flag when we return success from DDI
DxgkDdiNotifyAcpiEvent for dock event, so as to make sure OS is aware of
all the new displays attached, being driven on dock event.
|
|
BugID: 2171302,
2171303
|
DXVA DTM Fixes for Windows Vista*.
|
DVD
|
Windows Vista*, Windows*
XP
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Several DVA DTM failures on Windows Vista*, and needs some code cleanup
for both Windows Vista* and Windows* XP.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2170709,
2199989
|
Rotation does not work properly with dual view enabled.
|
Direct3D*, Rotation
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Rotation is not being considered in the SetDisplayMode function and
that is causing the issue. In this case, when we try to rotate, OS is
expecting the driver to take care of rotation and driver is creating extra
landscape surface. Then in the SetDisplayMode function, OS is sending the
handle for the original surface. Here, since this original surface is not
primary (extra landscape surface will be primary), we are failing this call
and this is resulting in the blank screen.
Fix Description: So, in this
SetDisplayMode function, check for orientation of the resource and if it
is not IDENTITY, then get the associated landscape handle and use that
handle. This is not happening with Single desktop with Aero as OS is taking
care of rotation by itself and no need of any logic tocreate extra surfaces
by the driver.
|
|
BugID: 2160055,
2223432
|
Selecting 1776x1000 clone mode (CRT and HDMI) causes
screen corruption on CRT with 8% underscan setting in INF file.
|
SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: The display EDID has 60 Hz based 1080i timings,
which results in dot clock to be calculated to a value = 74.40Mhz in PLL.
But the display supports 59.94 based timings for which the dot clock =
74.175, for that the customer needs to add a new DTD in the inf entry. This
will replace the current timing and will work for them. If we apply odd
valued X resolution on CRT, it causes corruption.
Fix Description: If the mode
that gets replaced via inf is a CE mode, then modify the corresponding
under scan mode only if it is added via VBT/inf. For all underscan
calculations, we make sure that X and Y resolutions are even-valued.
|
|
BugID: 2158369,
2221928
|
Extended desktop disabled after LID close/open suspend
operation.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Change in the
stream structure from initial one from MS. This is not a OS bug but a
change in stream. Its neither a bug fix nor a OS bug.
|
|
BugID: 2154658,
2174173
|
Failing Texture Stage Test.
|
Direct3D*
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® 915GM Express Chipset, Intel® G965 Express
Chipset, Intel® 945GM Express Chipset, Intel® GM965 Express Chipset, Intel®
915G Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: In several
cases, we needed to distinguish between the DX7, DX8 and DX9 interfaces,
and match the behavior of the reference rasterizer to pass DTM.
Fix Description: Fixed it appropiately.
|
|
BugID: 2142742,
2221906, 2287905
|
Gamma workaround in Windows Vista*.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Workaround
for Gamma implementation in Windows Vista*.
Fix Description: Maintain a status bit which gets set whenever
there is a CommitVidPN with at least 1 path OR SetSourceVisibility, with
visibility as TRUE. In UpdateActiveVidPnPresentPath, if this bit is set
then get gamma applied by CUI from registry, apply that and reset the bit.
If not set, apply gamma sent by OS or default as requested by OS. This bit
gets reset when CommitVidPn comes with no path or when SetSourceVisibility
gets called to turn off visibility. This issue is currently fixed in 32-bit
OS only.
|
|
BugID: 2141810,
2171304
|
System hangs and error message is displayed after
launching the 3D Mark 2001SE Benchmark.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root
Cause: The maximum primitive count and maximum vertex
index cap values were too big for DX runtime/3DMark2001 application to
handle.
Fix Description: Reduced maximum values slightly. Verified new values
are similar to other high end graphics cards, as they are not the maximum
theoretical values either.
|
|
BugID: 2128262,
2155763, 2222657
|
System has garbage during S4 resume.
|
SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® 915GM Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Turning off
PanelPower and Port during LVDS SetTiming call. Previously we used to keep
both the panel power and LVDS port ON during set timing. Because of this
certain panels, started to show garbage/corruption during mode changes. The
issue mentioned below is seen in Intel® GM965 Express Chipset platform due
to which S3/S4 operation are not happening. Panel is going off but system
not goes to S3 and S4. We can see through Platform LEDs that it shows S0
State.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2127504,
2255316
|
Grand Theft Auto San Andreas and FEAR hang.
|
Direct3D*
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The wrong
indices were being used when copying the vertex data into the HW vertex
buffer. This was happening because a copy and paste error that made into
the driver. Also. the SamplerState is not being copied into the command
buffer and eventually the batch buffer. The batch buffer still references a
sampler state, so the hardware is referencing stale data where the sampler
state indirect state should have been.
|
|
BugID: 2102022,
2255316
|
Implement AutoGenMipMap for Intel® 945G Express Chipset
and Intel® G965 Express Chipset for Windows Vista*.
|
Direct3D*
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Implement
AutoGenMipMap support for Intel® 945G Express Chipset and Intel® G965
Express Chipset Windows Vista*.
Fix Description: Fix
StretchRect failures seen from first ETM run and DTM GetRenderTargetData
failures from second ETM.
|
|
BugID: 2102022,
2255316
|
QueryProcessor break running under 64-bit OS.
|
Direct3D*
|
Windows Vista*-64
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: QueryProcessor
is generating an AV due to wrong cast in 64-bit environment.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2102022,
2255316
|
Adding Direct3D*/Ghal2D changes for NV12 Surface Format
Support.
|
Direct3D*, GHAL3D, GMM
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Adding
support for NV12 surface format on Windows Vista*.
|
|
BugID: 2102022,
2255316
|
GHAL3D updates and fixes.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: GHAL3D updates and fixes.
Fix Description: Change line antialias
alpha value from U0.4 to U0.8 when performing line antialiasing in the
pixel shader. Fix compute into mrf
assert with Gen5+. Switch threads in
system thread when writing certainarchitecture registers. Add new
SetDepthBufferCoordinateOffset client interface.
|
|
BugID: 2102022,
2198916
|
GHAL3D Default Value for VS Input Registers.
|
GHAL3D
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Insert
default value for VS input registers, if they are declared in shader, but
are not defined by Direct3D* or OpenGL. Also, vertex elements that are enabled,
but no declared as input registers, are ignored.
Fix Description: Added fix
for function inlining (CALL_CND) for VS and PS.
|
|
BugID: 2102022,
2198159, 2198390
|
UMD: Lock code with flags. GHAL2D Lock Discard on Vertex
Buffer. Get rid of flush per primitive in GHAL2D.
|
Direct3D*, DirectDraw*, DVD
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: These changes
are for some performance optimizations in GHAL2D.
Fix Description: 1) The
FlushBuffers function no longer takes a parameter. 2) The FlushBuffers function
on Windows Vista* would assume that Direct3D* owned the command buffer and
call Flush through GHAL3D. 3) Anywhere
DeviceContext->ActivateCommandBuffer( NULL ) was called to flush the
command buffer was converted to a DDI Flush call instead. 4) Removed flush
per primitive in C3DPixelShaderBlitter for both Windows Vista* and Windows*
XP. 5) Windows Vista* C3DPixelShaderBlitter now holds Lock addresses and
uses the aliasing (discardflag) on its vertex buffer. 6) Windows* XP
C3DPixelShaderBlitter also discards and allocates a new vertex buffer when
the existing one is full.
|
|
BugID: 2102022,
2197091, 2197203
|
Performance Optimization: BindPipeline Drity Flags.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Many improvements
made to the graphics driver for performance optimization.
Fix Description: Few major
changes are listed below: 1) SWState and HWState are now identical data
structures. 2) Added dirty flags for indirect states and state processor
context states. 3) Changed the inheritance model of the state processor
context. 4) Added function(s) to only set BindPipeline dirty flags at
BindPipeline time.
|
|
BugID: 2102022,
2173611, 2197091
|
F.E.A.R. performance benchmark has a temporary hang on
14.25.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Removed
unnecessary syncing of state with hardware.
|
|
BugID: 2102022,
2171299
|
Return values in UMD Flush() cause PRIMARYSURF.EXE to
fail.
|
Direct3D*
|
Windows Vista*
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fix for
return values in UMD Flush() that cause PRIMARYSURF.EXE to fail.
|
|
BugID: 2102022,
2159795, 2197091
|
DTM WDDM Present Stats Fix.
|
Direct3D*
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Some tests
were failing due to delay caused by deleting the immediate vertex and index
buffers after processing each primitive.
Fix Description: We now
delete the immediate buffers only when we run out of memory for
vertex/index buffers.
|
|
BugID: 2102022,
2143201
|
D3DCrossProcSharing.
|
Direct3D*
|
Windows Vista*
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The DTM test,
D3DCrossProcSharing, is asserting in the OS, because we setup our resource
flags incorrectly for a shared, non-lockable, volume texture.
Fix Description: The fix is
to make sure that if a resource is shared, we do not send the NonLockable
flag to GMM.
|
|
BugID: 2076950,
2171304
|
ValidateTSS and WireFrame Fixes for Windows Vista*.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fixes to
address random failures in the Lightmap.exe test and the wireframe test.
Fix Description: Fixed it appropiately.
|
|
BugID: 2076652,
2076658, 2157250
|
Enabling Blt's for Depth and Stencil Surfaces,
StretchRect failures, and NonPowTwo fixes.
|
Direct3D*, DirectDraw*
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: When doing
Sys->Vid and Vid->Sys blt's on Depth Buffer and Stencil Surfaces, the
source format can be different from the destination format. Previously the
driver would skip doing the blt in these cases. The SysMemPitch for
compressed surfaces should only be used for DX7.
Fix Description: Our blitting
functions need to pass the sub resource index of the surface used in the
blt. Also fixed the StretchRect test.
|
|
BugID: 205932,
2055212
|
A bug occurs because of missing overlapping BLT and
DepthFill support.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Bug occurs because
of missing overlapping BLT and DepthFill support.
Fix Description: Fixed it appropiately.
|
|
BugID: 2052319,
2254972
|
Corruption seen while running Foxbear or running WMP
clips.
|
DirectDraw*, DVD, GMM
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® G965 Express Chipset,
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: In 2D, 3D,
and Cube Map Texture Width should be 4 byte aligned. Also we need to make
sure Linear Planar Video Resources are 64 byte aligned.
Fix Description: Fixed it appropiately.
|
|
BugID: 2052319,
2199531
|
Wide Panel Modes Fix for Clone.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: When we try to
set a wide panel mode/unsupported mode on one of the displays in Clone from
OS property page, 8x6 mode gets set. This is because GDI relies on what
target modes we return for each pinned sourcemode. In this case, we will
not enumerate any target mode when the wide panel-like mode is chosen, so
OS sets 8x6 as a fallback error mode.
Fix Description: Fix is to enumerate the max supported
target mode for that display when the source mode is not supported directly
on CRTs and DFPs. This is to maintain consistency with the way we pin
target modes for such a source mode in RFV. The First cycle of ETM failed
since the issue still persists when the user tries to set12 x 8 resolution
from the OS properties page.
|
|
BugID: 2052319,
2198158, 2198394
|
Improve performance of blit commands through 3D pipeline.
|
DirectDraw*, GDI
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Fixed blts
and color fills with color keying to get rid of corruption. Fixed flush per command buffer to get rid
of slowness.
|
|
BugID: 2002879,
2255316
|
Fixing 15.1 regressions.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fixing
regressions in the 15.1 driver.
Fix Description: The CreateResource function was modifying
two input parameters which was causing some tests to fail. Changed the code
to work without doing this.
|
|
BugID: 22222651
|
DVD playback content flashing when HWA is enabled.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: VerticalLineStride
should be set back to 0 in procamp blt.
Fix Description: Fixed it appropiately.
|
|
BugID: 2294395
|
Issues caused because of back merge problem with IDDC and
Downscalar check-in.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fix issues
caused because of back merge problem with IDDC and Downscalar check-in.
|
|
BugID: 2292571
|
Hotplug detection does not work.
|
AIM
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root
Cause: When hot plug is enabled, it generates an
interrupt. But driver does not receive this interrupt as interrupt is not enabled
in the GMCH by this time. As a result interrupt state is not cleared in the
encoder and consequently it blocks all future hot plug interrupts.
Fix Description: Put a
workaround in UAIMEH. Added a new function UAIM_HotPlugInterruptClearWA()
in UAIMEHDFP.c, which gets called after enabling hotplug. In this function,
it is verified to see if there is any hotplug interrupt source at the
output corresponding to the pAIMObj. If there is an interrupt event at that
output, interrupt state is cleared.
|
|
BugID: 2292243
|
The resolution and color bit changes to the lowest
setting after switching output destination.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: With Windows*
XP driver on Windows Vista*, MinDispMode call was returning 6X4@8bpp
because the check for Windows Vista* OS was missing.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2291738
|
Audio on HDMI Panasonic receiver not enabled.
|
AIM
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root
Cause: ELD buffer size is greater than 32 with panel
connected. In UAIM_SetHDMIBufferData() there is "byBufSize >
32" condition check because of which ELD is not set. However with
SI1390 and CH7315 audio is working as there is some non-zero value in the
ELD buffer, but these values are invalid.
Fix Description: "byBufSize > 32" condition check is not
done in case of ELD data.
|
|
BugID: 2290648
|
VESA Modes blank out on HDMI Displays.
|
SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root
Cause: Scantype check was added in all the Mode Search
routines in SB. With these changes, when non-EDID modes say 10x7 was
applied on a HDMI Display with only 1080i support, no EDID timing could be
found and VESA timings were applied, which caused the display to blank out.
Fix Description: With the new changes, once a displays modes are added,
all the modes for which interlaced Timings will be applied, the
corresponding timing info is modified to indicate its interlaced. So when
10x7 is applied, RM knows an i timing is applied and sets the flag
appropriately and sends to SB.
|
|
BugID: 2289839
|
Stack overflow and underflow exception.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Intel® G965
Express Chipset and Intel® GM965 Express Chipset hardware no longer support
the mask stack exception.
Fix Description: Turns off
the enabling of the exception, adds code to manually trigger the exception,
and adds code to handle underflow exceptions (code already exists to handle
overflow exceptions).
|
|
BugID: 2289839
|
3DMark'06 Corruption Fix.
|
Direct3D*, DirectDraw*, GDI, GHAL3D, OpenGL*
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® G965 Express Chipset,
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Added code to
send flush command any time the render target or depth buffer changes.
Fix Description: Fixed it appropiately. Also fixes
corruption seen in 3DMark'06, Shader Model 3.0, and HDR Test 2, where the
sun moving across the screen leaves a noticable verticle white band.
|
|
BugID: 2288959
|
IN DDC Cursor Position are Mismatching while Rotating.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Same code
change from previous rotation issue.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2288780
|
Corruption and error message is displayed while playing
DVD movie.
|
DVD
|
Windows Media* Center, Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: WinDVD8 will
create larger surfaces for SD content, causing the condition for tiled temp
buffers to be triggered while ADI is enabled.
Fix Description: Instead, we
will tie use of Bob to tiled temp buffers since DI mode will be decided
during device creation and will not be changed throughout playback.
|
|
BugID: 2287887
|
Enabling OpenGL*.
|
OpenGL*
|
Windows* XP, Windows* XP-64, Windows* 20003-64, Windows
Vista*, Windows Vista*-64
|
Intel® Q33/Q35/G33 Express Chipsets-B, Intel® G965
Express Chipset
|
|
Resolution Description:
Root Cause: Enabling OpenGL*.
Fix Description: 1) Set Virtual
Address bits for Memory Interface commands. 2) Do GMM allocation for Tag
Page. 3) Check for Intel® Q33/Q35/G33 Express Chipsets Device ID. 4) Remove
GMM allocation for Tag page and use Miniport allocation. 5) Read additional
GMM mapped GTT offset for tag page if VTD is enabled and use that for Store
DWORD. 6) SetIndirectStateMemorySpace when state processor is created. (new
GHAL3D interface)
|
|
BugID: 2287739
|
When running clips named toy train there is massive green
corruption all over the player.
|
Direct3D*, DVD
|
Windows Vista*, Windows Vista*-64,
Windows Media* Center
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: There seems
to be a potential H/W issue with using Y-Tiled surfaces as the destination
when using VerticalLineStride (i.e. field decoding).
Fix Description: We will enable
X-Tiling instead until issue can be resolved.
|
|
BugID: 2255801
|
Incorrect resolution shown for LFP before entering
extended mode.
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: The modes
enumerated in LFP only and Monitor only modes are fine. But when we are in
LFP only mode and select Extended Desktop mode with CRT as primary and LFP
as secondary, the modes being enumerated for LFP in Notebook tab of CUI
increases from the max supported mode.
Fix
Description: Add
check for secondary device.
|
|
BugID: 2255316
|
Visual corruption in Serious Sam II with 32-bit index
buffers.
|
Direct3D*
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: When a 32-bit
index buffer followed a 16-bit index buffer, there was a rounding problem
when calculating the number of indices.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2255316
|
Fix Access Violation in the game code of GUN and FarCry
choppiness.
|
Direct3D*
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: A few performance
changes where checked-in which removed a flush per primitive. This caused
the driver to run out of command buffer space in a different place than
before(namely IssueQuery). This function did not flush when there is not
enough command buffer space.
Fix Description: Added code
to flush and then try to issue the query in this case.
|
|
BugID: 2254971
|
The color of mouse pointer becomes gray when setting
"Scheme" to "Windows Inverted" (system scheme).
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: In the
monochrome cursor of large size, if there is need to change the cursor info
with AND/XOR operations according to different mouse pointer schemes, it is
not possible with ARGB selection.
Fix Description: Changed the
ARGBcursor mode to XRGB cursor mode.
|
|
BugID: 2254912
|
OpenGL Driver Stability Drop 1.
|
OpenGL*
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® Q33/Q35/G33 Express Chipsets-B, Intel® G965
Express Chipset, Intel® GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: OpenGL* is
merging a fully tested timestamp of the OpenGL* private branch.
Fix Description: The
following changes are included: A0 enabling. Quake IV crash fix for low
memory setups .
|
|
BugID: 2254550
|
Pixel Repetition occurs on VGA, 480p, and 576p when multi
channel audio is supported.
|
SoftBIOS
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Pixel
replication can be set to 2X if 1440x480p@59.94/60Hz (VIC 14,15) and
1440x576p@50Hz (VIC 29,30) formats are supported by the display. Today,
graphics driver is not supporting these CEA formats.
Fix Description: Current logic
for setting pixel replication should be modified to check if the display
supports pixel replication (i.e. supports CEA formats 1440x480p and
1440x576p) before setting the pixel repetition factor. Support for setting
pixel replication factor to 2X can be added when support for CEA formats
with VIC 14,15 and 29,30 is added to the graphics driver.
|
|
BugID: 2254488
|
Bordering on CRT cannot be removed in monitor only mode.
|
SoftBIOS
|
Windows* 2000, Windows* XP, Windows Media* Center,
Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Fixed a bug in
the logic for calculation of the scaling options.
|
|
BugID: 2223707
|
The Menu on Device Manager shows Generic Non-PnP Monitor
for LVDS LCD.
|
Miniport, SoftBIOS
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: SB is
returning the bFakeEDID flag properly in the case of SDVO LVDS. This is
getting checked in Miniport and hence the issue. Actually this is expected
as of the current architecture. There is a bug in the case of the current
Integrated LVDS behavior due to which the bFakeEDID flag was not getting
updated properly. That was the reason why this issue is not seen in the
Integrated LVDS display. Actually Windows Vista* expects EDID if the
display is LVDS and doesnt care about whether this is faked or not. so, now
the fix is to take care of the issue with the bFakeEDID flag in SB in the
case of Integrated LVDS.
Fix Description: In Miniport,
we will not consider this flag if the display is LVDS.
|
|
BugID: 2223432
|
1080i SetInputTimingPart2 opcode incorrect setting.
|
AIM
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: If we apply odd valued X resolution on CRT, it
causes corruption.
Fix Description: Now for all underscan calculations, we make sure that
X and Y resolutions are even-valued.
|
|
BugID: 2223308
|
The blue screen is displayed and system hangs after
opening IGT page then closing the lid switch.
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: This Issue was happening due to uninitialized
variables in CuicomEscape.c
Fix
Description: Initialized the variables in CuicomEscape.c
|
|
BugID: 2223147
|
When system boots, noise appear after progress bar of
Windows Vista*.
|
SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: During Post, VBIOS uses High Res Plane to display
OEM Logos. This data in the frame buffer is not cleared and hence the corruption.
SB do not program Start Address as part of setmode. There are cases that
Start Address is not available during setmode in Windows Vista*. Windows
Vista* OS will call Driver to BlankVideo before setmode and enables back
the video after setmode through SetSourceVisibilty DDI. But SB internally
disables and enables back the video as part of setmode. This causes the
corruption. This was required for Windows* XP to synchronize with render.
The same is not required in Windows Vista*.
Fix Description: Added a new flag bVideoBlankingReqd in setmode
interface. Clients have to set/reset this flag based on their need.
Windows* XP setmode path has to set this flag and Windows Vista* setmode
path has to reset this flag. We need to set this flag in all cases within
SB where we use SetMode Interface.
|
|
BugID: 2222704
|
Mouse Pointer shadow is broken.
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset, Intel® GM965 Express
Chipset
|
|
Resolution Description:
Root Cause: When we
resume from monitor off, DxgkDdiSetPointerShape() is called with
MASKEDCOLOR_CURSOR and then DxgkDdiSetPointerPosition() then
DxgkDdiSetPointerShape() is called for ALPHA_CURSOR and then
DxgkDdiSetPointerPosition() is caled. EnableCursor() is not called for
ALPHA_CURSOR in DxgkDdiSetPointerPosition() because the cursorinfo
visibility flag is set to true for MASKEDCOLOR_CURSOR and it remains true
for ALPHA_CURSOR. The cursor control bits are not changed for the
ALPHA_CURSOR type due to not calling enablecursor results into corruption.
Fix Description:
DxgkDdiSetPointerShape() checks if the cursor type is different from
previous cursor type, and then disable the cursor and set its visibility
flag to false.
|
|
BugID: 2222533
|
EDID Read Failure Issue during S3/S4 Events.
|
SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Currently we
have synchronization in AIMENCODER_ServiceHotPlugInterrupt Function where
we get a Lock on GAL before doing detection.
Fix Description: Added the acquire
GAL and Release GAL Interfaces before we do a detection.
|
|
BugID: 2222158
|
Support for ARGB or XRGB output in media CSC kernels.
|
DVD
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Windows Vista* expects the post-processing module to handle
both XRGB (32bpp) and ARGB (32bpp) destination surfaces. When the output is
ARGB, the Alpha channel must be set to 0xFF; if the output is XRGB, the
Alpha (X) channel must be set to 0x00.
Fix Description: The driver
must be able to recognize the output format, and pass the information for
the kernel. The CSC kernel must be able to set the alpha channel based on
data receive from the driver.
|
|
BugID: 2222087
|
BugCheck THREAD_STUCK_IN_DEVICE_DRIVER.
|
DirectDraw*, DVD
|
Windows Vista*
|
Intel® 915GM Express Chipsetm
|
|
Resolution Description:
Root
Cause: The issue is due to the Surface Tag not being
hooked to the surface structure after HW source copy in DdUnlock() call.
That may cause graphics memory access order issues.
Fix Description: Added tag to the surface structure so that SW can
synch with HW that the surface is finished being written to from previous
operation.
|
|
BugID: 2221989
|
Both 30Hz Interlaced and 60Hz are shown for 1080i
supported HDMI displays.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: HDMI display
with Max Capability of 1080i is to show refresh rate of 60i and 60p in OS
Property Page and CUI. This is because of enumerating 1920x1080x60Hz GTF
mode. This is a special case.
Fix Description: Do not report modes greater than Max Capability
of the Monitor.
|
|
BugID: 2221928
|
Resume from S4 blanks out on TV.
|
CUI COM
|
Windows Vista*
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Removing
current config check for CUICOM.
|
|
BugID: 2211572
|
Rotation Radio buttons of 90 degree and 180 degree are
not hidden.
|
CUI2
|
Windows* XP, Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Rotation Radio buttons of 90 degree and
180 degree are not hidden.
Fix Description: Added additional method CustomizeRotation()
to hide the controls based on customization options mentioned in registry.
|
|
BugID: 2211552
|
Laptop does not play DVDs and reports the wrong display resolution
and refresh rate information when using Component and LVDS in Clone Mode
display.
|
Miniport
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: GfxGetVidPnDetails never returned mode information as
per OPM expectation from type GET_INFO_FROM_PIPE_TYPE. The requirements of
OPM are different. They need Source X,Y, Target Frequency expressed as a
Rational Number (Num/Denom format) and Target Scan line format.
Fix Description: Introduced a new enum type for OPM
[GET_OPM_INFO_FROM_PIPE_ID] wherein given a Pipe ID [passed in CPUpdate
function after mode set], we populate Source X, SourceY, Target Freq Num,
Target Freq denom, Target ID and Target Interleave format.
|
|
BugID: 2211338
|
Unnecessary symbols or characters in CUI > Information
> Video BIOS.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: bsrtversion
holds zero memory which was causing the issue.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2211334
|
Fn+F3 does not work after [A driver stopped responding
and has recoverd] message is displayed.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Once
TDR happens, driver does a Graphics Device Reset with which all the MMIO
and I/O register gets initialized to default values. Due to which
ActiveToggle list table index and GR18 I/O register were also getting
zeroed out as a result of which Display switch was blocked.
Fix
Description: So now once a TDR happens we save the state of
scratch registers before a graphics reset and restore the relevant fields
after the reset is done.
|
|
BugID: 2211310
|
Visual corruption when using the DirectX PIX tool to
render single frame captures.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: When
disabling render targets only disable the Z-Buffer if all possible render
targets are disabled.
|
|
BugID: 2211106
|
Age of Empire 3 game shows a black display after menu
setting.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*, Windows Vista*-64
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Removes
unnecessary MOV instructions for macro expansions.
Fix
Description: This prevents pixel shaders
from being generated that contain more instructions than the hardware
supports.
|
|
BugID: 2211055
|
DXVA Memory Allocation / Deallocation Enhancements.
|
DVD
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Several memory-related
enhancements have been made.
Fix
Description: For Windows Vista*, add a
wrapper function for malloc() that will memset() the allocated memory to 0.
Memory Allocation / Deallocation has been enhanced for both Windows Vista*
and Windows* XP. We currently do not zero out newly allocated memory on
Windows Vista*, which causes Access Violations later.
|
|
BugID: 2210986
|
DXVA2: API Invalid cases: 3.15.3, 3.15.9, and 3.15.10
failing/
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The case for D3DDDICAPS_FILTERPROPERTYRANGE
is not handled in umdcaps.cpp, causing 3.15.3 to fail. 3.15.9 and 3.15.10
have been root-caused as DTM bugs.
Fix
Description: Add support for
DxvaGetFilterPropertyRange. Also re-write dxvacaps_lh.c.
|
|
BugID: 2210941
|
Cannot change display destination by Hotkey after
hibernation.
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: On resume from S4, display switching is initialized to
OFF in customer BIOS. System BIOS expects an explicit call to enable
display switching.
Fix Description: On resume from S4, evaluate _DOS to enable display
switching.
|
|
BugID: 2200749
|
LCD timing becomes 40Hz in 60Hz setting.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: The specified system has a panel that supports 40 and
59.968. Windows Vista* will list 59 and 60 Hz in this case due to some
changes that they have made to address the anti-aliasing issue. In case of
CLONE, CUI gets the mode list from driver, which does not have 59 Hz at
all.
Fix Description: Fallback mechanism inRFVN to select the target mode
when user preferred values are not found was modified. Earlier we used to
revert back to the native timing with the least EDID supported RR. This
change, in this specific scenario will make the driver revert back to 60Hz.
Even with this change, OS would still return 59Hz to CUI as the current RR.
Since 59 is not present in the mode list returned by the driver, they used
to fallback to zero-th element in the RR array which will be 40. Sufficient
changes made in CUI such that in case of such mismatches, they will read
the RR from registry. Changes have been made in miniport to update the
DISPLAY1_MonitorXXX values to the primary RR in registry in case of CLONE. Also
contains some small cleanup in VidRegistry.c to use safe string functions
instead of normal C routines.
|
|
BugID: 2200741
|
TDR issue during Aspect Ratio Option setting.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: The OS was missing some VBLANK notification during
the setting up of compensation in the driver and this was the casue of the
TDR. This was happening since the driver was doing a SetTiming call while
doing the SetCompensation call.
Fix Description: The fix mainly separates the SetTiming call from the
SetCompenstaion call. During the SetCompensation, only the value of the
corresponding varaible is changed in driver. It does not take effect until
the next SetMode call comes in. CUI has been modified to do a SetMode call
after issuing the SetCompensation call to the driver.
|
|
BugID: 2200734
|
Production Issue with interlaced VC-1 content.
|
DVD
|
Windows Vista*, Windows Vista*-64,
Windows Media* Center
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Production Issue with interlaced VC-1
content.
Fix Description: Return S_FALSE=1 instead of FALSE=0.
|
|
BugID: 2200307
|
Vendor Specific Infoframes Support for SDVO HDMI.
|
AIM, CUI2, CUISDK
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Add support
for Vendor Specific Infoframes for SDVO HDMI.
Fix Description: An
application can send Vendor Specific Infoframe data using a CUI SDK API.
This data will then be sent to the SDVO encoder by the graphics driver, and
transmission of VS infoframe will be enabled.
|
|
BugID: 2200004
|
DVD player closes itself without playback.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: The kernel
table is different for Intel® GM965 Express Chipset.
Fix Description: Use the
correct kernel table.
|
|
BugID: 2200001
|
LCD/DVI Rate always 50/60Hz after S3/S4 in Clone mode.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: This issue is
reproduced only with Multi RR panels. Dotclock for all centered modes which
we add for LVDS was not correct. Because of this GetRRFromTiming Interface
used to return wrong RR. Dotclock for 50Hz and 60Hz were same. For a panel
with 16x12 Native Resolution, we add all modes in the static table and
replace just the refresh rate. This is fine in case of Windows* XP where
OS/Miniport dont care about timing. In case of Windows Vista*, 1024x768
x50Hz and 1024x768 x 60Hz has the same Dotclock. Because of this
GetRRFromTiming interface fails.
Fix Description: Modified the
Dotclock with the following formula : Dotclock = HTotal X VTotal X RR.
|
|
BugID: 2199988
|
Fix for caching of MapArray, SamplerArray, and PixelShader States.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*, Windows Vista*-64
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: There was a problem
with the caching of the MapArray, SamplerArray, and PixelShader state.
Fix Description: The fix was
to use SafeAcquire to increase the reference count of the corresponding
StateCacheData once the data was created, and SafeRelease on a Delete call
or FreeLRU call.
|
|
BugID: 2199450
|
Free CP memory during driver unload.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® Q33/Q35/G33 Express Chipsets-B, Intel® G965 Express
Chipset, Intel® G965 Express Chipset, Intel® 945GM Express Chipset, Intel®
GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: Memory allocated by content protection was not
freed during driver unload.
Fix Description: Free the memory by hooking into content protection
during driver unload.
|
|
BugID: 2199246
|
No DTD in the (.INF) file of the Graphics Dr, Windows
Vista*64bit.
|
INF
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Missing DTD registry
entry in the INF.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2198918
|
Windows* XP Driver on Windows Vista*
Fix.
|
OpenGL*
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: Did not support the case for a Windows Vista* OS
with Windows* XP driver.
Fix Description: Added support for running the Windows* XP driver on
the Windows Vista* OS.
|
|
BugID: 2198916
|
OpenGL* Windows Vista*-64 bit driver.
|
OpenGL*
|
Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: Lack of three registry key settings in the INF.
Fix Description: Added the correct registry key settings to the INF.
|
|
BugID: 2198760
|
Snowy screen after renew events.
|
DVD
|
Windows Vista*, Windows Vista*-64,
Windows* 20003, Windows Media* Center, Windows* XP, Windows* XP-64
|
All Platforms
|
|
Resolution Description:
Root Cause: The driver
used to try to restore the HDCP state after a monitor only power event.
This code was erroneously being executed during system power events as
well.
Fix Description: Removes all
code related to restoring HDCP state after power events, which in the end
should have no effect to real world applications. The only effect will be
on test application that do not poll. This also fixes an issue in which
there was an error reporting the connector types for non-HDCP displays.
|
|
BugID: 2197727
|
Unexpected display device is enabled instead of expected
display device if you reboot system with "Multiple Display" mode.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Driver is not making any recommendation for the configuration
when OS comes up with an RVT (RecommendVidPnTopology) with request reason
DXGK_RVT_INITIALIZATION_LKGOVERRIDE on boot time. OS could use the last
known topology LFP+DFP. But immediately after this, it comes up with
DXGK_RVT_AUGMENTATION_LKGOVERRIDE where we were recommending CRT as
secondary (This is a request reason which is also used to recommend a
secondary when going to extended desktop from OS properties page), since at
this point we are not aware that it is happening as a part of boot sequence
and CRT is the highest priority device after LFP.
Fix Description: Since we cannot differentiate when we get the later
RVT call, as to whether it is boot time or going to extended desktop from
OS properties page. Also It is legal to go with OSs LKG secondary even for
extended desktop from OS properties page. We are now checking for the
request reason, and if the OS has an LKG, then we are not recommending
anything as secondary. In the other case, it is still OK to recommend a
secondary from drivers side.
|
|
BugID: 2197706
|
MV Related Changes and addition of four modes for SDTV.
|
TVout
|
Windows Vista*, Windows Vista*-64,
Windows* 2000, Windows Media* Center, Windows* XP, Windows* XP-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Code changes needed.
Fix Description: 1) Code
Changes for four new modes on SDTV. 2) Code Changes for MV Compliance on
Windows Vista*. 3) Code Changes required
for Synchronization between Driver and CUI when TV Standard is changed.
|
|
BugID: 2197631
|
INF Customizations are not working in latest driver
build.
|
CUI2
|
Windows Vista*-64, Windows*
XP, Windows* XP-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: For disabling
the tray icon, only HKCU entries were read.
Fix Description: If no value is
present in HKCU, HKLM should be read.
|
|
BugID: 2197203
|
Dynamic CURBE Allocation.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: CURBE design using fixed, pre-defined data structure.
Fix Description: Changed the
CURBE design to use a dynamically allocated buffer, instead of the fixed,
pre-defined data structure.
|
|
BugID: 2197091
|
Limiting the number of allocation indices per command
buffer.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: These changes
prevent GHAL3D from creating more allocation indices than the OS expects.
The KMD defines the number of allocation indices and passes this to the OS
and the UMD. These changes call the USERMODE_DEVICE_CONTEXT (in Direct3D*)
and BackEndCmdTransport (in OpenGL*) classes to get the size of the
allocation table.
Fix Description: When
requesting an allocation index, GHAL3D now checks that it is not larger than
the largest index in the allocation table. If it is, then it will return an
index of 0 which is an disabled index and set the RETVAL.Success to false
and RETVAL.FlushCommandBuffer to true.
|
|
BugID: 2196455
|
Cannot change device from CRT 12x10 to LCD only with CUI
SDK TestApp Tool.
|
Miniport
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Issue is happening because App created using CUI
SDK sends a mode set without querying whether the mode is supported on
the LFP display or not. While
switching from CRT (12x10) (which
support 12x10 does not since panel max is 14x9), Miniport gets a
RFVN(USER_MODE), where we recommend 12x10 on LFP (both source/target mode),
but it is not supported on bug. LFP
Post recommendation, OS calls EVCM we prune out both 12x10 in source/target
modeset since it is not supported on 14x9 panel (panning case).
Fix Description: Need to prune this mode(12x10) in RFVN call and
fallback to native resolution of panel (i.e. 14x9) which is supported on
the panel.
|
|
BugID: 2196442
|
Add COPP_ConnectorType_Internal to LVDS connectors.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset, Intel® GM965 Express
Chipset
|
|
Resolution Description:
Root Cause: There is a COPP
flag that should be returned when an app queries the connector type and the
type is LVDS.
Fix Description: This flag
indicates that the connector is an "internal" connector and does
not require protection.
|
|
BugID: 2196129
|
Unable to set 900x700 OEM resolution when with CRT.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: For EDID displays,
the OS does not take all the modes that are returned by SB. In the case of
EDIDless displays, Miniport calls GetMonitorModeList from
RecommendMonitorModes function, which will report all the modes from SB and
hence the issue is not happening there.
Fix Description: The fix
involves moving the logic inside the RMM function in MP to
GetMonitorModeList in SB. Here, SB will check whether there is any OEM
custom mode in the mode table and if there are some, then SB will be
returning the entire mode list even in the case of EDID displays.
|
|
BugID: 2196103
|
WDK EVR DCT Player Multimon.
|
DVD
|
Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root
Cause: When running a 32-bit test, since UMD Loaded is
32-bit, it needs to be copied out of the WOW64 directory.
Fix Description: Install the 32-bit UMD to the wow64 directory and make
sure that the kernels are loaded from the correct location.
|
|
BugID: 2194832
|
When HDMI display support 1920x1080i@50Hz and 60Hz both.
50Hz cannot be selected.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Interlaced
was not taken care for Rounding RR issue.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2194297
|
IGT page hangs after setting LFP+CRT EXTD display mode
and 90 Degree rotation mode for CRT device twice.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: In SetCursysconfig
all the calls to CDS for secondary inactive (90/270) fails, hence need to
add a workaround for rotating the mode and making a CDS call.
Fix Description: Workaround
added.
|
|
BugID: 2194260
|
Hot Plug is not functioning properly for DVI and HDMI.
|
AIM
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: A detailed investigation
of this issue shows that its purely a user error in this case. It happens
because as soon as you touch the connector pins to the connector you get an
interrupt and perform a device detection. (Note: At this moment the
connector is not fully plugged in) So by putting the delay, we have some
time before the user fully plugs in the connector, and then we read the
EDID and other stuff.
|
|
BugID: 2173373
|
Tile support in DXVA.
|
DVD
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: There are some issues in kernel for linear surface
support.
Fix Description: Add tile support in video processing.
|
|
BugID: 2172628
|
Rotation related fixes.
|
Rotation
|
Windows Vista*, Windows Vista*-64,
Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Changes are
done to fix the Enable/Disable Rotation through CUI page.
Fix Description: Driver will not
be looking into the registry key Display1_EnableRotation which will be
used by CUI. So, driver will enumerate rotation angles without based on
Enable/Disable Rotation key in CUI. When rotation is disabled through CUI,
driver still enumerates rotation orientations. But CUI will not give the
rotation options. This is same as Windows* XP parity.
|
|
BugID: 2172524
|
DVI-I configuration is not correct.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root
Cause: Even if the Function 1 of IGD is disabled in CMOS
driver was reporting 2 sources as supported to OS.
Fix Description: Before reporting the Number of supported sources to OS
well check is Device2 Function1 is disabled in CMOS. If function 1 is
disabled then well report 1 view to OS.
|
|
BugID: 2170909
|
"Adapter String" is "Intel(R) G945 Express
Chipset Family" in Display Settings.
|
Miniport
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: [Windows
Vista*32/Intel® G965 Express Chipset/15.0] "Adapter String" is
"Intel(R) G945 Express Chipset Family" in "Display
Settings".
Fix Description: Fixed it
appropiately.
|
|
BugID: 2170708
|
Rotation Modifications to incorporate peer review comments
and address InkBall bug.
|
Direct3D*, GMM, KMD Render
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Incorporating
the peer review comments received for the UMD rotation changes done by the India
team for enabling 180 degree hardware rotation on Intel® G965 Express
Chipset.
Fix Description: Fixing
InkBall bug related to rotation.
|
|
BugID: 2170683
|
FBC stride will come from RM instead of GMM. Fence ID
will come from GMM.
|
|
|
|
|
Resolution Description:
Root Cause: PC will no
longer query GMM for the stride for FBC. The stride will be passed through
RM and then to PC. PC will still get the Fence ID from GMM.
Fix Description: This
includes fixes where Miniport calls grmnotifyevent when aspect ratio
options are changed. This will trigger RM to get the latest dotclock and
provide it to PC to recalculate its SR watermarks. Fixes screen jitters
seen in maintain aspect ratio modes. Includes a fix where Intel® GM965
Express Chipset GFX SW CapID was being read from the wrong place.
|
|
BugID: 2170339
|
Playback MPG2 has frame drop on MP11.
|
DVD
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Use the same
size as Windows* XP for residual difference buffer.
|
|
BugID: 2160063
|
Bank screen is displayed after unplugging HDMI TV and
plugging EFP then resuming from Suspend state (S3).
|
AIM
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: On Windows
Vista*, SetPowerState call for AIM encoders never happens when system goes in
sleep state; as a result AIM encoders remain in D0 state. This is due to LH
sends us DEVICE_STANDBY for sleep operation unlike Windows* XP (which sends
DEVICE_OFF). Currently AIM encoder does not support DEVICE_STANDBY which is
mapped to D1 power state.
Fix Description: To handle
this condition, we will continue calling UAIM_SetEncoderPowerState from
SoftBios and handle D1/D2 cases within AIM.
|
|
BugID: 2159067
|
Cannot check F/W version of SDVO device on CUI.
|
CUI2
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: This was
happening with binaries built from VS2003 where the general setting for igfxTray.rc
was C/C++ Compiler tool, where as this should be Resource compiler tool.
Fix Description: So this has
been fixed with VS2005 which was checked in recently.
|
|
BugID: 2158274
|
Enable Video Process Constriction Support for Windows Vista*.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Constriction
support is a Logo Requirement and is required to pass DTM Tests.
Fix Description: This fix
adds Video Processing Constriction Support.
|
|
BugID: 2158274
|
Add IMC3 support as decode RT.
|
DVD
|
Windows Vista*
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Presently
IMC4 planar format is being used as the decode rendertarget. There are
fundamental problems for tiled IMC4 surfaces due to tile alignment for the
V plane. It is possible to use IMC3 on all Windows Vista* platforms as it
is already being used for the current platform. Tiled surfaces are
preferred for performance reasons.
Fix Description: Direct3D* team will implement tiled
support for planar surfaces, so this DCN will add the support for IMC3
surfaces for DXVA code. In addition, enabling tiled temp buffers for post-processing
for scenarios not using adaptive deinterlacing.
|
|
BugID: 2158249
|
Black screen is displayed on CRT after resume from the
suspend state (S3).
|
Miniport
|
Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: During S3/S4
resume, OS sends commits on extended desktop topologies, with paths
originating from non-affected source IDs. And before the in-active paths is
applied, a cleanup is performed, which leaves the resources attached to
non-affected source ID as active. This causes a lot of confusion in RM for
Intel® G965 Express Chipset, and it returns invalid PLANE/PIPE assignments
for futher mode sets. This used to work properly for LKP/CAL platforms.
Fix Description: The fix now
identifies topologies with paths originating from non-affected source ID in
commit, and forces it back to SPSD config.
|
|
BugID: 2158020
|
W - DCT 5.3 QFE - PC99TA 6.8.2 Color Key Support -
BSODxEA.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® GM965 Express Chipset, Intel® 915G Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: a) When a
timing is applied, pixel doubling is set based on the actual timing
programmed versus the OS mode size. For 400x300 running on a panel @
800x600. However, there are displays which support timings which are less
than 8x6. (ie. 640x400) In such cases, pixel doubling should not be set. b)
Since pixel doubling is set based on the timing, upper layers (miniport
etc) do not have any way to find if its running on low res mode. Cursor
positioning for eg, required this info to double the cursor co-ordinates
when display plane is doubled. Hence, defined an interface to get the
timinginfo as well as lowres mode.
Fix Description: Fixed it
appropiately.
|
|
BugID: 2156963
|
Extended desktop not retained after closing/opening LID.
|
ACPI
|
Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: OS is not relying upon the graphics driver
reporting the target is attached with the system via
CbIndicateChildStatus(ChildUid, Attached). OS current implementation is not
proper where DMM relies on monitor driver to tell them of the device being
attached. They are targeting a clean-up in this area post Windows Vista*.
Fix Description: It was recommended to set the Polled flag along
withdisplay switch flag in ACPI lid event handling
|
|
BugID: 2156742
|
TMM/Windows Vista* bug fixes.
|
CUI
|
Windows Vista*
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: Overlay is not POR for 15.0.
Fix Description: Used customization in INF to remove the Overlay
button, so no change is needed in code. When the feature is enabled, we
just remove the INF entry for removing Overlay button.
|
|
BugID: 2156620
|
945GM+Si1390 HDMI, centered "not natively supported
mode" do not have correct VideoTiming/InfoFrame.
|
AIM, SoftBIOS
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Currently for non-CE modes, VIC is set to 0 since
that mode is not found in Short Video Descriptor table. Moreover aspect
ratio is being calculated using the X and Y resolutions of the selected
mode, which is wrong.
Fix Description: Currently for non-CE modes, VIC is determined based on
mode in which it is centered. Each VIC has a corresponding aspect ratio as
mentioned in the CEA spec. Now we are setting aspect ratio corresponding to
the VIC.
|
|
BugID: 2155563
|
CxSR flicker when resuming from S3/S4.
|
Power Conservation
|
Windows* 2000, Windows* XP, Tablet Pc, Win2003, Windows
Media* Center, Windows* XP-64, Win2003-64, Windows Vista*, Windows Vista*-64
|
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Added
save/restore of chicken bit register for Intel® GM965 Express Chipset
Windows* XP/Windows Vista*.
Fix Description: This makes sure
that proper chicken bit configuration is maintained through S3/S4.
|
|
BugID: 2155501
|
Cursor ghost.
|
SoftBIOS
|
Windows Vista*
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: We used to
see white color patch around the cursor when the desktop comes up. This is
because gamma correction bit for cursor was not enabled.
Fix Description: Fix is to
enable/disable gamma correction while enabling/disabling cursor.
|
|
BugID: 2154804
|
Garbage/unwanted characters seen under CUI information
page->Monitor tab.
|
CUI2
|
Windows* XP, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: This is
because the value of registry key is different in Windows* XP and Windows
Vista*.
Fix Description: Fixed it appropiately.
|
|
BugID: 2143201
|
Code Optimization for Hardware Vertex buffer and Hardware
Index Buffer to remove the happening of Device lost.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: 1) Optimize
some functions in d3dindex.cpp (i.e. AllocateMemory, AliasMemory, SetStride
and SetMaxIndices). 2) Optimize some functions in d3dvertex.cpp (i.e.
AllocateMemory, AliasMemory, SetStride and SetMaxVertices). 3) Optimize
GetMaxVertices in Gen4VertexProcessor.cpp.
|
|
BugID: 2102022
|
Pixel Shader Compiler Infrastructure.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows Vista*, Windows Vista*-64,
Windows* XP, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Adds
infrastructure to the pixel shader compiler.
|
|
BugID: 2102022
|
Conversion to Barycentric clipping to avoid register
overflow.
|
Direct3D*, GHAL3D, OpenGL*
|
Windows* XP, Windows Vista*
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Restucturing
of Clipper Kernel Program generation code for maintainability, accuracy, and
performance benefits. Clipping cycle clips from input to output buffer,
then moves output back to input for next clip plane. Input is sized as
three (maximum input to clipper is a triangle) but clipping can generate up
to six total vertices (worst case). So moving these six back into input
would overwrite whatever is after the input buffer. Currently, there are
three unused registers after the input buffer, so this issue would only be
exposed when there are a large number of vertex elements. Adding these
extra register introduced another issue by which we are running out of
available registers to hold all the clipped vertex data.
Fix Description: The solution
is to use Barycentric coordinates during clipping and generate the clipped
vertices afterwards one at a time.
|
|
BugID: 2102022
|
FFT&L and Vertex Shader Fixes.
|
Direct3D*
|
Windows* 2000, Windows* XP, Windows* 20003, Windows
Media* Center, Windows* XP-64, Windows* 20003-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: 1) Fixed to
vertex shader register mapping for FFT&L. 2) Fixed and enabled LOCALVIEWER
lighting. 3) Enabled ViewportZClipping ( Fix PointSprite DCT Test ). 4)
HWVP Mode Registry Indicator. 5) PS/VS Constant Buffer Optimization. 6)
Removed Tabs ( Coding Standard ). 7) Added registry key for PUREDEVICE (
for debug purposes only - feature not implemented ). 8) Fixed memory leak
in destructor of CD3dVertexShader - deleted GHAL3D shader. 9) Fixed Vertex
Shader output declaration issue with usageindex. 10) Enabled DX9 Vertex
Decl Types and UBTYTE4 for DX8 ( UBYTE4 needed for game compatibility per
BattleField2 ). 11) Disabled Constant Buffer Copy Optimization for Pixel
Shaders due to failure observed running 3D Mark'06 Deep Freeze. 12) Added
additional check for bEnableHardwareVertexIndexBuffer in addition to device
support before enabling caps. 13) Added fix for wireframe-TNL failure in
HWVP mode. 14) Updated with \main.
|
|
BugID: 2100064
|
Uninstaller Error message is displayed after installing the
graphics driver via have disk method and then uninstalling it from
Installed programs.
|
INF
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® GM965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Windows Vista* appears to have yet another difference from
all the other versions of windows that breaks the method currently used by
the co-installer to retrieve the source path of the INF when installing.
The preferred method to get this path from a co-installer is to include the
DIRID 1 in a custom section in the INF that the co-installer can read.
|
|
BugID: 2077456
|
Support planar surfaces in UMD ColorFill.
|
Direct3D*
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The UMD
entrypoint ColorFill is missing support for Planar surfaces.
Fix Description: This change
adds that support.
|
|
BugID: 2076909
|
DTM failures on GetRasterStatus Test and some associated clean-ups
in MMDMM code.
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Fix for DTM
failures for GetRasterStatus for Main.
Fix Description: 1) Optimizations
in Miniport code to read the current active topology. Added a new function
to return the current active topology pointer, instead of copying all the
data, to be used in places where no updation of the active topology is
required. 2) Change in SB to compare the current scanline value with
VActive value to identify VBLANKs. 3) Change in SB to return the current
VActive value, so that Miniport can scale down the scan line value in cases
where scaling is applied. 4) Using GrmValidateDisplayConfig instead of
GrmQueryDispConfig from IsSupportedVidPn so that GRM will not be caching
any queried configurations. 5) Caching of the OS provided timing values
instead of the actual timings set by SB in the current config while
extending the desktop with aspect scaling set, so we do not provide wrong
timing values to GRM. Currently, if we are on LFP SPSD with aspect scaling
/ centering set, we will store the actual timing that is set by SB. While
extending the desktop, OS would now do a modeset on the second source only
which would result in us using the cached timing value to GRM. This might
be custom timing values, which would then be failed by GRM.
|
|
BugID: 2076902
|
Gamma Correction DTM Fix.
|
Direct3D*
|
Windows Vista*
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Turning on
the D3DPMISCCAPS_POSTBLENDSRGBCONVERT CAP bit to enable post blend SRGB
conversions on Windows Vista*.
|
|
BugID: 2015284
|
The VBIOS version is not displayed in the Information
Page of CUI.
|
SoftBIOS
|
Windows* XP, Windows* XP-64, Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: For Platforms
not supporting Int10, SMI and ACPI Opregion, we need to get BIOS version
from VBT.
Fix Description: For Platforms
not supporting Int10, SMI and ACPI Opregion, we need to get BIOS version
from VBT.
|
|
Issues Resolved
in Build 1111
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
|
BugID: 2199245
|
Driver will not install with test certificate signature
files due to timeout
|
Installer,
INF
|
Windows Vista*,
Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
|
Root
Cause: Missing quotes under
"igfxColn_v1101.dll ...". Along with this error in the setuplog
the iGfxMFCoInstaller function did not exist. Additionally, Vista changes the method that was used by the co-installer
to retrieve the source path of the INF when installing. The preferred
method to get this path from a co-installer is to include the DIRID 1 in a
custom section in the INF that the co-installer can read.
|
|
|
BugID: 627160,
627161
|
OGL 64 bit fix for 15.0
|
OGL
|
Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
|
Root Cause: Enables three registry key settings that are required
for the OGL 64 bit driver. Fixes the drawing rectangle size we send to
GHAL3D when DWM is enabled so that we can blit in full-screen.
|
|
|
BugID: 628160,
628161
|
3D Perf optimization, GHAL3D Optimization: Enable IZ for
Vista-Lakeport
|
D3D
|
Windows Vista*,
Windows Vista*-64
|
Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
|
Root Cause: Gen3 DMA buffer size for performance set to optimal 64K.
Calistoga corruption fix when presenting to primary when running 3D apps
without Aero
|
|
|
Issues
Resolved in Build 1106
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
|
BugID: 2196087
|
Media kernels with RGB32 output must have alpha channel set
to 0xff instead of 0x00.
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
|
Resolution Description:
Root
Cause: Color Space Conversion media kernels were setting the
alpha channel to 0x00 instead of 0xff, which is eWindows* XPected behaviour
(same as Gen3).
Fix Description: Fixed the color space conversion kernels to set the
alpha channel to 0xff.
|
|
|
BugID: 2102022, 2198159,
2198390
|
Fix for UMD: Lock code with flags. GHAL2D Lock Discard on
Vertex Buffer. Get rid of flush per primitive in GHAL2D.
|
D3D, DD, DVD
|
Windows* XP, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: These changes
are for some performance optimizations in GHAL2D. A lot of code in D3D is
touched in this change but at low risk as it is mainly an interface change.
The highest risk changes are in GHAL2D where the vertex buffer code has
changed significantly. Here are the changes: 1) The FlushBuffers function
no longer takes a parameter. 2) The FlushBuffers function on Vista would assume that D3D owned the command buffer and
call Flush through GHAL3D. 3) Anywhere
DeviceContext->ActivateCommandBuffer( NULL ) was called to flush the
command buffer was converted to a DDI Flush call instead. 4) Removed flush
per primitive in C3DPixelShaderBlitter for both Vista
and WINDOWS* XP. 5) Vista C3DPixelShaderBlitter now holds Lock addresses
and uses the aliasing (discardflag) on its vertex buffer. 6) WINDOWS* XP
C3DPixelShaderBlitter also discards and allocates a new vertex buffer when
the existing one is full.
|
|
|
BugID: 2076902
|
Vista : Gamma
Correction Caps fix.
|
D3D
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: Subtests in
the Gamma Correction DTM test were failing on LKG, CAL,
and BWG because a new cap, D3DPMISCCAPS_POSTBLENDSRGBCONVERT, was added to the
DX9L refract. This cap enables post blend SRGB conversions.
|
|
|
BugID: 2197203
|
INT_15_0 fixes avalon Glyph Test in Vista
by rounding texture addresses when min and mag filter mode is not nearest.
|
D3D, DD, GHAL3D, KMD, OGL
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
|
Resolution Description:
Root Cause: 1) When min or
mag filter mode is not nearest neighbor, set bits in sampler state to round
texel addresses. 2) Change KMD Render to offset all vertices by -0.5f.
|
|
|
BugID: 2197240,
2197276
|
MSFT:WDDM: Intel:M400: Restoring External Only results in
black screen on both displays. (MS bug#1862456)
|
Miniport
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
|
Resolution Description:
Root Cause: In Vista, Miniport does not call SoftBIOS to set Primary
Pipe for VBIOS [Bit 10 of 0x71414]. Therefore when a user boots with LFP
only, VBIOS sets this bit to indicate Pipe B [a value of 1]. On every mode
setcall, we call SetDisplayDevice with current pipe and plane and in the
process SMI 050D gets called with the associated Port-Pipe Matrix. The Pipe
Index for this is read from Bit 10 of 0x71414. Since this is set wrongly,
the corresponding Bits indicating CRT does not get set for the SMI. Toshiba
BIOS has an embedded controller that'll block display from showing contents
until this SMI uses right values. Hence, display could not be seen on CRT
only. They do not have a similar logic for Int LVDS or DVI. Fix was to Call
SB before SetDisplayDevice so that the SMI has right values for Port-Pipe
matrix.
|
|
|
BugID: 2196442
|
Add COPP_ConnectorType_Internal to LVDS connectors on
Vista (INT_15_0).
|
DVD
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset, Intel® GM965 Express
Chipset
|
|
|
Resolution Description:
Root Cause: There is a
COPP flag that should be returned when an app queries the connector type
and the type is LVDS. This flag indicates that the connector is an
"internal" connector and does not require protection.
|
|
|
Issues
Resolved in Build 1101
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
BugID: 2173373
|
Tile support in DXVA.
|
DVD
|
Windows Vista*
|
Intel® G965 Express Chipset
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: There are some issues in kernel for linear surface
support.
Fix Description: Add tile support in video processing.
|
|
BugID: 2102022,
2126973, 2143201, 2197091, 2197203
|
Windows Vista* Bug Fix:
D3DCrossProcSharing.
|
D3D
|
Windows Vista*
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The DTM test
D3DCrossProcSharing is asserting in the OS because we setup our resource
flags incorrectly for a shared, non-lockable, volume texture. The fix is to
make sure that if a resource is shared, we do not send the NonLockable flag
to GMM.
|
|
BugID: 2015284
|
In Windows Vista*
OS, the VBIOS version is not displayed in the Information Page of CUI.
|
SoftBIOS
|
Windows* XP
Windows* XP-64,
Windows Vista*
Windows Vista*-64
|
Intel® G965 Express Chipset
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: For Platforms
not supporting Int10, SMI and ACPI Opregion, we need to get BIOS version
from VBT. (E.g.: Windows Vista* on Intel® 945G Express Chipset and Intel®
945GM Express Chipset.)
|
|
BugID: 2196103
|
Windows Vista*64 DTM Fixes - WDK EVR DCT Player Multimon.
(32-bit)
|
DVD
|
Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root
Cause: When running a 32-bit test, since UMD Loaded is 32-bit,
it needs to be copied out of the WOW64 directory.
Fix Description: Install the 32-bit UMD to the wow64 directory and make
sure that the kernels are loaded from the correct location.
|
|
BugID: 2197091
|
Windows Vista*:
Limiting the number of allocation indices per command buffer. (PC 15.0)
|
D3D, GHAL3D, OGL
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: These changes
prevent GHAL3D from creating more allocation indices than the OS expects.
The KMD defines the number of allocation indices and passes this to the OS
and the UMD. These changes call the USERMODE_DEVICE_CONTEXT (in D3D) and
BackEndCmdTransport (in OGL) classes to get the size of the allocation
table. When requesting an allocation index, GHAL3D now checks that it is
not larger than the largest index in the allocation table. If it is, then
it will return an index of 0 which is an disabled index and set the
RETVAL.Success to false and RETVAL.FlushCommandBuffer to true.
|
|
BugID: 2194832
|
When HDMI display support 1920x1080i@50Hz and 60Hz both.
50Hz cannot be selected.
|
SoftBIOS
|
Windows Vista*, Windows Vista*-64
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Interlaced
was not taken care for Rounding RR issue.
|
|
BugID: 2194260
|
Hot Plug is not functioning properly for DVI and HDMI.
|
AIM
|
Windows Vista*, Windows Vista*-64
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: A detailed
investigation of this issue shows that its purely a user error in this
case. It happens because as soon as you touch the connector pins to the
connector you get an interrupt and perform a device detection. (Note: At
this moment the connector is not fully plugged in) So by putting the delay,
we have some time before the user fully plugs in the connector, and then we
read the EDID and other stuff.
|
|
Description:
Resolution
Root Cause: It was
discovered that the BUFFER_INFO instruction on Gen3 platforms requires
QWORD alignment for the address when using 8-bit linear texture. This code
change implements the workaround.
|
|
Issues
Resolved in Build 1096
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
BugID: 2102022,
2143201
|
Microsoft® Windows Vista*
Bug Fix: D3DCrossProcSharing.
|
D3D
|
Microsoft® Windows Vista*
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The DTM test,
D3DCrossProcSharing, is asserting in the OS, because we setup our resource
flags incorrectly for a shared, non-lockable, volume texture. The fix is to
make sure that if a resource is shared, we do not send the NonLockable flag
to GMM.
|
|
BugID: 2158274
|
MPG2 Playback has frame drop on MP11.
|
DVD
|
Microsoft® Windows Vista*
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Use the same
size as WINDOWS* XP for residual difference buffer.
|
|
BugID: 2076909
|
Fix for DTM failures on GetRasterStatus Test.
|
Miniport
|
Microsoft® Windows Vista*,
Microsoft® Windows Vista*-64
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: Fix for DTM
failures on GetRasterStatus test. This is a temporary fix to address the
GetRasterStatus test. The ideal fix would involve changes in both MP and SB
and would required more effort, and there by would only be targeted post
PV. Interim solution is to read the registers and return back the values
from MP.
|
|
BugID: 2102022,
2171299
|
Fix for return values in UMD Flush() that cause
PRIMARYSURF.EXE to fail.
|
D3D
|
Microsoft® Windows Vista*
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fix for return
values in UMD Flush() that cause PRIMARYSURF.EXE to fail.
|
|
Issues
Resolved in Build 1092
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
ID: 93736,
91613
|
3rd Party players, Windows Media Player 11 -
movie lags when playing DVD movie.
|
N/A
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Kernel post processing code was not enabled in the
driver yet
|
|
BugID: 2171302,
2171303
|
DXVA DTM Fixes for Windows Microsoft® Windows Vista**.
|
DVD
|
Windows Microsoft® Windows Vista**,
Windows* XP
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: This DCN
fixes several DXVA DTM failures on Windows Microsoft® Windows Vista**, and
also does some code cleanup for both Windows Microsoft® Windows Vista** and
Windows* XP.
|
|
BugID: 2076950,
2171304
|
ValidateTSS and WireFrame Fixes for Windows Microsoft®
Windows Vista**.
|
D3D
|
Windows Microsoft® Windows Vista**, Windows Microsoft®
Windows Vista**-64, Windows* XP, Windows*
XP-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Fixes to
address random failures in the Lightmap.exe test and the wireframe test in
BWG Windows Microsoft® Windows Vista**.
|
|
BugID: 2076652, 2076658,
2157250
|
Windows Microsoft® Windows Vista** Bug fixes: Enabling
Blt's for Depth and Stencil Surfaces, StretchRect failures, and NonPowTwo
fixes.
|
D3D, DD
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset, Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: When doing
Sys->Vid and Vid->Sys blt's on Depth Buffer and Stencil Surfaces, the
source format can be different from the destination format. Previously the
driver would skip doing the blt in these cases. The SysMemPitch for
compressed surfaces should only be used for DX7. Intel® 945G Express
Chipset and Intel® 945GM Express Chipset still have 5 or 6 failures, but
Intel® G965 Express Chipset passes 100%. Also fixed the StretchRect test in
this DCN. Our blitting functions need to pass the sub resource index of the
surface used in the blt.
|
|
BugID: 2172628
|
Rotation related fixes.
|
Rotation
|
Windows Microsoft® Windows Vista**, Windows Microsoft®
Windows Vista**-64, Windows* XP-64
|
Intel® G965 Express Chipset, Intel® 945GM Express
Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Changes are
done to fix the Enable/Disable Rotation through CUI page. Driver will not
be looking into the registry key Display1_EnableRotation which will be used
by CUI. So, driver will enumerate rotation angles without based on
Enable/Disable Rotation key in CUI. When rotation is disabled through CUI,
driver still enumerates rotation orientations. But CUI will not give the
rotation options. This is same as Windows* XP parity.
|
|
BugID: 2170708
|
Windows Microsoft® Windows Vista** Rotation Modifications
to incorporate peer review comments and address InkBall bug.
|
D3D, GMM, KMD Render
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Incorporating
the peer review comments received for the UMD rotation changes done by the India
team for enabling 180 degree hardware rotation on BWG. Fixing InkBall bug
related to rotation.
|
|
BugID: 2156742
|
TMM/Windows Microsoft® Windows Vista** bug fixes.
|
CUI
|
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: Overlay is not POR for 15.0.
Fix Description: Used customization in INF to remove the Overlay
button, so no change is needed in code, we just remove the INF entry for
removing Overlay button.
|
|
BugID: 2155563
|
Fix CxSR flicker issue when resuming from S3/S4.
|
Power Conservation
|
Win2000, Windows* XP, Tablet Pc, Win2003, Windows Media*
Center, Windows* XP-64, Win2003-64, Windows Microsoft® Windows Vista**,
Windows Microsoft® Windows Vista**-64
|
Intel® GM965 Express Chipset
|
|
Resolution Description:
Root Cause: Added
save/restore for Intel® GM965 Express Chipset Windows* XP/Windows
Microsoft® Windows Vista**. This makes sure that proper configuration is
maintained through S3/S4.
|
|
BugID: 2154804
|
Garbage/unwanted characters are seen under CUI
information page->Monitor tab.
|
CUI2
|
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root Cause: This is
because the value of registry key is different in Windows* XP and Windows
Microsoft® Windows Vista**. In Windows Microsoft® Windows Vista** the key
looks like "@monitor.inf,%pnpmonitor%;SONY CPD-G420". Out of
which SONY CPD-G420 needs to be displayed on CUI.
|
|
Issues
Resolved in Build 1087
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
|
BugID: 2141810,
2171304
|
Hang / D3D - Test ID 8445 - 3D Mark 2001SE - system hangs
and error message is displayed after launching the 3D Mark 2001SE
Benchmark.
|
D3D
|
Windows Microsoft® Windows Vista**, Windows Microsoft®
Windows Vista**-64, Windows* XP, Windows*
XP-64
|
Broadwater-G, Intel® GM965 Express Chipset
|
|
|
Resolution Description:
Root Cause: Root Cause: The
maximum primitive count and maximum vertex index cap values were too big
for DX runtime/3DMark2001 application to handle. Fix Given: Reduced maximum
values slightly. Verified new values are similar to other high end graphics
cards, as they are not the maximum theoretical values either.
|
|
|
BugID: 205932,
2055212
|
A bug occurs because of missing overlapping BLT and
DepthFill support.
|
D3D
|
Windows Microsoft® Windows Vista**, Windows Microsoft®
Windows Vista**-64, Windows* XP, Windows*
XP-64
|
Intel® G965 Express Chipset, Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: Bug occurs
because of missing overlapping BLT and DepthFill support.
|
|
|
BugID: 2170909
|
[Microsoft® Windows Vista*32/BWG/15.0] "Adapter
String" is "Intel® G945 Express Chipset Family" in
"Display Settings".
|
Miniport
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
|
|
|
Resolution Description:
Root Cause: [Microsoft®
Windows Vista*32/BWG/15.0] "Adapter String" is "Intel® G945
Express Chipset Family" in "Display Settings".
|
|
|
BugID: 2158249
|
Black screen is display on CRT after resume from the
suspend state (S3).
|
Miniport
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
|
|
|
Resolution Description:
Root Cause: During S3/S4
resume, OS sends commits on extended desktop topologies, with paths
originating from non-affected source IDs. And before the in-active paths is
applied, a cleanup is performed, which leaves the resources attached to
non-affected source ID as active. This causes a lot of confusion in RM for
Broadwater, and it returns invalid PLANE/PIPE assignments for futher mode
sets. This used to work properly for LKP/CAL platforms. The fix now
identifies topologies with paths originating from non-affected source ID in
commit, and forces it back to SPSD config.
|
|
|
BugID: 2155501
|
Cursor ghost.
|
SoftBIOS
|
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: We used to
see white color patch around the cursor when the desktop comes up. This is
because gamma correction bit for cursor was not enabled. Fix is to
enable/disable gamma correction while enabling/disabling cursor. This
change does not have any impact on XP.
|
|
|
Issues
Resolved in Build 1078
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
|
BugID: 2143201
|
Multiple Windows Microsoft® Windows Vista**
Performance Changes for D3D.
|
D3D
|
Windows Microsoft® Windows Vista**-64
|
Intel® G965 Express Chipset,
Intel® 945GM Express Chipset,
Intel® GM965 Express Chipset,
Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: Multiple
performance enhancements for Windows Microsoft® Windows Vista** will be added
with this code change: 1) Optimize vertex buffer copies re-enabling. 2)
Holding the VB and IB locks. 3) Lock Discard (Aliasing). 4) Holding the
querybuffer lock. 5) Cacheable query buffer. 6) Larger scratch BLT buffer
size to 1 MB.
|
|
|
BugID: 2143551
|
Fix snowy screen on DVI hot plug.
|
AIM
|
Windows Media* Center,
Windows* 2003, Windows* 2003-64, Windows* XP, Windows*
XP-64, Windows Microsoft® Windows Vista**, Windows Microsoft® Windows Vista**-64
|
All Platforms
|
|
|
Resolution Description:
Root
Cause: HDCP encryption was left on because during the hot
plug, calls to UAIM were failing and immediately returning.
Fix Description: Whenever there is an error condition in UAIM, HDCP
should be turned off, along with returning the error.
|
|
|
BugID: 2158344
|
Fix for TV jitter issue from CGMSA.
|
TVOUT
|
Windows Media* Center,
Windows* 2003, Windows* 2003-64, Windows* XP, Windows*
XP-64, Windows Microsoft® Windows Vista**, Windows Microsoft® Windows Vista**-64
|
Intel® 945GM Express Chipset,
Intel® GM965 Express Chipset
|
|
|
Resolution Description:
Root
Cause: Improper waiting for ready status for integrated TV
out CGMS-A registers.
Fix Description: Correct the wait condition from 1 to 0.
|
|
|
BugID: 2156137
|
CUI Hotkey disable register cannot work.
|
CUI2
|
Windows Microsoft® Windows Vista**,
Windows* XP
|
Intel® 945GM Express Chipset
|
|
|
Resolution Description:
Root Cause: The registry reading
procedure was not proper. Modified the registry reading procedure, so that
it will check for HKCU entry. If HKCU entry is not there, then check for
HKLM entry. If HKLM entry also is not there, then we use the default keys.
|
|
|
BugID: 2157990
|
Display Expansion setting is still on Full Screen(No
Border) after applying the scheme with Maintain Aspect Ratio setting.
|
CUI2
|
Windows Microsoft® Windows Vista**, Windows Microsoft®
Windows Vista**-64
|
Intel® G965 Express Chipset,
Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root
Cause: The values of the pBuilder structure was not copied
to m_bckBuilder.
Fix Description: The data of pBuilder structure is copied to
m_bckBuilder.
|
|
|
Issues
Resolved in Build 1070
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
|
BugID: 2102022
|
Windows* Microsoft® Windows Vista* Bug Fix: AlphaBld.exe
failures.
|
D3D, DD
|
Windows Microsoft® Windows Vista**,
Windows* XP
|
Intel® G965 Express Chipset,
Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root Cause: The driver is
not tagging the rendertagets that are a flip chain correctly which causes the
driver to skip flushing the command buffer when dependencies existed. Code
change is to store the sub resource index in the multiple render target
list. This variable is now retrieved from the list instead of being derived
whenever it is needed.
|
|
|
BugID: 2101956
|
MaskedColor Cursor support.
|
MINIPORT
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset,
Intel® 945GM Express Chipset,
Intel® 945G Express Chipset
|
|
|
Resolution Description:
Root
Cause: Certain games are using MaskedColor type of cursor
which was not supported and used to fall back to SW cursor. This SW is resulting
in cursor flicker and corruption in the games because of some OS issue.
Fix Description: Added changes for MaskedColor cursor support through
HW cursor.
|
|
|
BugID: 2102022,
2126965, 2157970
|
Fix DTM MipGen.exe (Auto Generation of MipMaps) in Intel®
945G Express Chipset and Intel® G965 Express Chipset Windows Microsoft® Windows Vista**.
|
D3D
|
Windows Microsoft® Windows Vista**, Windows Microsoft® Windows Vista**-64, Windows* XP
|
Intel® G965 Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: For Intel®
G965 Express Chipset Windows Microsoft® Windows Vista**: fix is in SetRenderTarget()
to get the x and y offsets of the drawing rectangle for cube resources.
This applies only to Gen4 Windows Microsoft® Windows Vista**. For LKP
Microsoft® Windows Vista*: fix is in SetupCubeTextureCoordinates() where we
were incorrectly calculating the +X,+Y, -Z cube texture coordinates. I
noticed that these three faces have something in common: a negative axis
(ie., +X face has a z axis, +Y face a z axis, and Z face a x axis). So
I negate the sign of the texture coordinate that is positioned in the
negative axis, and the test passed 100%.
|
|
|
BugID: 2076910
|
Windows Microsoft® Windows Vista** Bug Fix: Vertex and
Index Buffers, CrossProc, and miscellaneous DirectDraw tests. Small YVU9
surfaces.
|
D3D, DD
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset,
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: We were
returning the wrong failure code. When
out of memory we need to return DDERR_OUTOFVIDEOMEMORY.
|
|
|
Issues
Resolved in Build 1065
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
BugID: 2128662
|
Intel® GM965 Express Chipset] CUI Information page has
noise.
|
SOFTBIOS
|
Windows* XP
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
Intel® 945GM Express Chipset
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Panel fitter
is not required when in native or panning mode. Recently issues were reported
on quality difference between panel fitting enabled/disabled when in native
mode. The immediate fix for this is to make sure panel fitter is freed up.
1) During set timing check for source == destination. If so, enabling
scaling is not required and set ucStretching flag to 0 while calling GAL.
2) When we received a call to get current compensation, check for native
timings, and if so, return failure. This will make sure CUI will not show
the scaling options if in native/panning mode. This fix is required, so
that we will not end up showing options but unable to apply them. With
Intel® GM965 Express Chipset coming up, sDVO HDMI dynamic scaling need to
be tested on it when LVDS is active. Freeing up scaler will make sure HDMI
will get the scaler for downscaling activity.
|
|
|
BugID: 2102022
|
Bundled fixes for Windows Microsoft® Windows Vista**:
Chroma Key Enabling, StateBlock Recording, Ghal2D Saturate Fix (TextureFilter--DTM),
Ghal2D Shader Fix for Gen3, and Format Fix for YVU9.
|
D3D
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: Bundled fixes
for Windows Microsoft® Windows Vista**: Chroma Key Enabling, StateBlock
Recording, Ghal2D Saturate Fix (for the DTM test TextureFilter), Ghal2D
Shader Fix for Gen3, and a D3D Format Fix for YVU9 which was being checked
improperly in d3dresource.
|
|
|
BugID: 2140732
|
Cannot playback AVI file with sound on MP11.
|
DVD
|
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Lacking of RGB as source support.
Fix Description: Add RGB support as source in Post Processing.
|
|
|
BugID: 2140397
|
DVD cannot play on Windows Microsoft® Windows Vista** / Intel®
945GM Express Chipset.
|
DVD
|
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
|
|
Resolution Description:
Root
Cause: Invalid buffer format was reported when VLD is
enabled in SKU.
Fix Description: Rework the buffer format reporting for both Windows*
XP and Windows Microsoft® Windows Vista** to report required buffer formats
only.
|
|
|
BugID: 2127714
|
Blank screen is displayed on CRT after unplugging the
secondary device from DDC display
mode and resuming from Suspend State (S3).
|
MINIPORT
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
Resolution Description:
Root
Cause: An OS Bug. OS supplies a disconnected device to
drive and does not drive display to other connected device.
Fix Description: In Commit, we ignore monitor connectivity check even
if asked by OS. This will result in a commit being notified to the EM. EM
will simply try to find a driver default configuration given the current
configuration. Now if OS was attempting to drive on a disconnected device,
EM will take care and drive to a connected display. Otherwise it will abort
doing nothing. The mechanism was always there in the code, it has just been
enabled by setting a flag in resume handler.
|
|
|
BugID: 2102022
|
Fix for setting EarlyDepthTestEnable for Intel® 945G
Express Chipset.
|
D3D, GHAL3D
|
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: On Intel®
945G Express Chipset, there are two EarlyDepthTestEnable bits depending on
whether the driver is doing zone rendering or classic rendering. The bit
for zone rendering was being set when it needs to be set for classic
rendering for Windows Microsoft® Windows Vista** Intel® 945G Express
Chipset. Also changed how we set the build number that the UMD sends to the
KMD in the command buffer header.
|
|
|
Issues Resolved
in Build 1028
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project(s)
|
|
BugID: 2126948
|
BWG fix for BSOD on BWG G965 system.
|
MINIPORT
|
Windows Microsoft® Windows Vista**
|
Intel® G965 Express Chipset
|
|
Resolution Description:
Root Cause: Added missing
BWG sku changes as part of DCN 459656 to fix the issue.
|
|
|
BugID: 2103380
|
BSOD fix while going to S3 in Windows Microsoft® Windows
Vista**.
|
MINIPORT
|
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset, Intel® 945G Express Chipset
|
|
Resolution Description:
Root Cause: The OS sends
hPrimaryAllocation as NULL in Commit. In such a case, Miniport should not
invoke GMM with this parameter eg GmmIsTiled interface. GmmGetLockPitch call
was protected earlier. When this parameter is NULL, Miniport should assume
memory as Linear
|
|
|
Issues
Resolved in Build 1014
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project's)
|
|
|
Bug
ID: 2041584
|
900x700
resolution does not list in Display Settings page and IGT page after
flashing the modified BIOS
|
SoftBIOS
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
|
|
|
Resolution
Description:
Root Cause: The issue was
happening since the timing for the mode (900 * 700) given in the DTD, as
part of the reproduction data, was wrong. The mode corresponding to the
given DTD was 904 * 700. If this mode (904 * 700) is made in the VBT
instead of the 900 * 700 for the given DTD, the issue is not happening and
the 904 * 700 mode is appearing in the CUI as well as the OS property
pages. Hence this issue has been resolved as a test bug.
Fix Description:
|
|
|
Bug
ID: 2028502
|
Resuming
from Standby returns to a garbled/unusable screen
|
ACPI/Power
Management
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: OS bug on Microsoft® Windows Vista* builds
prior to 5431.
Fix Description: No driver fix needed.
|
|
|
Bug
ID: 2052942
|
The
blank playback is displayed while playing the MPEG movie file via Media
Player 11 application.
|
DVD
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: Due to the KMD limitation, the width of
residual difference buffer was adjusted to the size of buffer and the height
was adjusted to 1. Therefore, sw codec copy the data for one macroblock
only.
Fix Description: 1. KMD implements the special handling
for MC_Buffer, and UMD set the special flag without doing the adjusment.
2. Set drawing rectangle always in post processing due to
the usage model change in which D3D and DXVA share the same HW context.
|
|
|
Bug
ID: 2055239
|
The
Question message is displayed while installing the graphics driver from Start\
Run\setup.exe with command -f2c:\Installerlog\mylog.txt.
|
Installer
(Stand
Alone)
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: The -f2 switch has been changed in
baseline 14.21 and greater to require a space between the switch and the
path to closer match command line syntax standards.
(ie
setup.exe -f2 c:\Installerlog\mylog.txt)
Fix Description:
|
|
|
Bug
ID: 2078162
|
When
change aspect scaling Mouse cursor changes it side and its not visible
format.
|
SoftBIOS
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: When we Disable Cursor Plane and enable it
back during Settiming we did not program Cursor Base Address properly. In CURSORPLANEBASE_Enable function we
did not pass Cursor Base address.
Fix Description: CURSORPLANEBASE_Enable
needs to get baseaddress and then call EnableCUrsor function.
|
|
|
Bug
ID: 2076161
|
When HD video
content is played with Windows Media player11 only blank video play is
observed.
|
DVD
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
|
|
|
Resolution
Description:
Root Cause: No temp buffer is allocated with YV12
format.
Fix Description: Allocate temp
buffer for all planar formats we can support.
|
|
|
Bug
ID: 2100531
|
Cursor
not visible after display power test
|
NT-Miniport
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
Cause: The OS
does not always send a SetPointerPosition with Visible Flag set as FALSE before
doing a mode change. During a mode set, SB disables the cursor plane but
the private flag is not reset. After a mode change, SetPointerPosition
looks at the CursorVisible private flag to decide whether to enable cursor
to cursor not being seen sometimes.
Fix
Description: Fix
is to reset the private flag after a successful mode set, so that in the
subsequent SetPointerPosition call, cursor plane is always enabled.
|
|
|
Bug
ID: 2053878
|
ACCESS_VIOLATION_igdumd32!DxvaCreateVideoProcessDevice+61
|
DVD
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
Cause: WMP calls
to allocate resource of DecodeRenderTarget with YUY2 format, driver fails
in GMM because this is an invalid format of DecodeRenderTartet.
Fix
Description: Add error checking in UmCreateResrouce to
check the invalid condition and reject the call if invalid.
|
|
|
Bug
ID: 2076167
|
ACCESS_VIOLATION_igdumd32!D3d::D3DRESOURCE::IsLocked+c
|
DVD
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
Cause: The
pointer to a D3DResource is used after the resource is destroyed.
Fix
Description:
Reset the pointers to NULL in DXVA reference list and destination pointer
when the D3DResource is destroyed.
|
|
|
Bug
ID: 2100535
|
AV for pHwDev->bRotation_enable =
FALSE
|
Rotation
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
Cause: Cursorcachebuffer
will be created (for rotating the cursor shape) only when rotation is
enabled. In SetCursorShape call,
this cache buffer is being accessed even if it isNULL and resulting in
access violation.
Fix
Description:
Check for Cursorcachebuffer and access that only if it is not NULL. This
will not affect the cursor in non rotated mode and avoids the access
violation.
|
|
|
Issues
Resolved in Build 1007
|
|
|
Reference
No.
|
Description
|
Affected
Component(s)
|
Affected
OS(s)
|
Affected
Project's)
|
|
|
Bug
ID: 2028490
|
Bugcheck
encountered when running PnPStop tests
|
SoftBIOS
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: Free all the allocated memory while unloading the driver.
Assign the variable to NULL after Successful release.
Fix Description: Fixed
appropiately.
|
|
|
Bug
ID: 2028494
|
TDR
encountered while running Miscellaneous DirectDraw* tests
|
KMD
Render
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: OS Bug in build 5355
Fix Description: [1] OS Build 5341 [2] Code fix in __GmmGen3AcquireSwizzleRange
( ) to limit the total number of display surfaces that can be locked at one
time, to 6. On the 7th attempt to lock, GMM will fail the Swizzle request
forcing OS to release a previously swizzled surface to make room.
|
|
|
Bug
ID: 2028499
|
TDR
Encountered while running LDDM_CancelFlip test
|
KMD
Render
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: OS Bug in build 5355
Fix Description: [1] OS Build 5341 [2] Code fix in
__GmmGen3AcquireSwizzleRange ( ) to limit the total number of display surfaces
that can be locked at one time, to 6. On the 7th attempt to lock, GMM will
fail the Swizzle request forcing OS to release a previously swizzled
surface to make room.
|
|
|
Bug
ID: 2028503
|
WLP 2.2 RGB
Rasterization test hangs
|
SoftBIOS
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: There were multiple issues with Planar
surfaces not being displayed properly when planar->RGB conversions are
being used in Gen3 Microsoft® Windows Vista*. Certain Planar surfaces were having their
U & V Planes created with the wrong widths and heights. The Planar to RGB shader needed 3
distinct sampler states in Gen3 (as opposed to 1 for Gen4). The Sampler States were only ever being
set for the 1 sampler state.
Fix Description: Fixed appropiately.
|
|
|
Bug
ID: 2053422
|
Blank
screen is displayed while playing DVD Annex 3.0 DVD movie with WinDVD7
Application
|
D3D
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: There were multiple issues with Planar
surfaces not being displayed properly when planar->RGB conversions are
being used in Gen3 Microsoft® Windows Vista*. Certain Planar surfaces were having their
U & V Planes created with the wrong widths and heights. The Planar to RGB shader needed 3
distinct sampler states in Gen3 (as opposed to 1 for Gen4). The Sampler States were only ever being
set for the 1 sampler state.
Fix Description: Fixed appropiately.
|
|
|
Bug
ID: 2053437
|
Movie
stops but audio stream plays normally in some sigutations with progressive
DVD content
|
D3D
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root Cause: There were multiple issues with Planar
surfaces not being displayed properly when planar->RGB conversions are
being used in Gen3 Microsoft® Windows Vista*. Certain Planar surfaces were having their
U & V Planes created with the wrong widths and heights. The Planar to RGB shader needed 3 distinct
sampler states in Gen3 (as opposed to 1 for Gen4). The Sampler States were only ever being
set for the 1 sampler state.
Fix Description: Fixed appropiately.
|
|
|
Bug
ID: 2052625
|
Blank
selection is displayed on video Standard Option with TV display mode
|
CUI
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
Cause: For SDTV
also the video format selection was changed depending on the screen
resolution.
Fix
Description: Put
a check for HDTV and the format selection was done only for HDTV.
|
|
|
Bug
ID: 2015335
|
EFP
Refresh Rate can not be changed to 75hz after setting display mode from CRT
at 75Hz to CRT+EFP DDC mode
|
NT -
Miniport
|
Microsoft®
Windows Microsoft® Windows Vista**
|
Intel® 945GM Express Chipset
Intel® 945G Express Chipset
|
|
|
Resolution
Description:
Root
cause: CUI reads registry
Display2Monitor_(displayUID) to get the refresh rate of secondary in
clone. This registry is not updated by
the driver after setting mode in DxgkDdiCommitVidPn()
Fix: Updated the registry Dispaly2Monitor_(display_UID) for
clone with the refresh rate of secondary.
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|