1. Color layer has a destination rectangle, RGBA color and plane
alpha. There is no gralloc buffer. Layer flag HWC_COLOR_FILL
denotes color layer. MDP BLIT happens via MDP_SOLID_FILL mode.
2. There is no color member in HWC layer, so RGBA color value is
passed via "tranform" member from framework to HAL.
3. Update HWC query to enable framework to query for Color layer
support in HAL, at run-time.
Change-Id: I2c698007c1689779fe86d549093bb5285432dc5f
This changes removes the use of 5551 and 4444 pixel formats
as its not supported anymore
Change-Id: I817510c3c7c0c3d95924c2ffc79e5727a6d6c48f
(cherry picked from commit 1fbd4a9662b07d53a3f220c91b3e85ca571a4938)
On HDMI connect or other use cases where finish_copybit
is called, reset dst_surface_mapped and dst_surface_base
as destination surface is already unmapped.
Change-Id: I0073a46178e6b1ecfdbf0a08d8790af4271bc37b
CRs-fixed: 481685
CRs-fixed: 502555
Since mdp composition is now an async call, the waiting for fences
is now no longer needed inside copybit. In addition, how
composition ioctl is called is slightly different as a result
needing to pass in aquire fences device performing compsition
Change-Id: Ia12dfb2960ba2fc78b14e776984ffe0c3fe45fdb
Add check in copybit for Max. layers supported by HWC. Increase
max. RGB surfaces to MAX_NUM_LAYERS to support known use cases.
CRs-Fixed: 481226
Change-Id: Ie30ab0ef3047a304133cb547902ae11fecc5db22
- Check MAX_SURFACES while mapping GPU addr
--- While mapping gpu address, first check if the request
is less than the MAX_SURFACES allowed.
--- Without this check, when the request exceeded MAX_SURFACES,
there was memory leak as that gpu addr was not unmapped
- Map the destination surface only once
--- The destination surface for the c2d blit will be the
FRAME_BUFFER_TARGET per transaction
--- Map this only once from the copybit_clear which is the first
call by hwc for a transaction
- Increase MAX_RGB_SURFACES to 12
--- To handle use cases where there are more than 8 RGB layers
Change-Id: Iffacde30ee88b9c21d0895cb1922aa981bb079e9
CRs-fixed: 475327
- Each client API exposed by the copybit C2D needs
to be protected.
- Add protection for clear_copybit and blit_copybit
- This is needed as clear_copybit was overriding the
mapped gpuaddr while copybitwaitthread was waiting
for transaction to finish and unmap
- After waitforTimestamp, the new addresses were
unmapped leading to GPU Page fault
Change-Id: Ieeac10e73b05d3b47e602b2de4b4e64892737c9c
CRs-fixed: 470430
Instead of assuming the premultiplied alpha based on the color
format, use the blending parameter passed by the HWC.
Change-Id: Ib494644b6f9c4bf13d8ea28327ccc2058a52aae8
CRs-fixed: 466732
Remove usage of C2D_NO_BILINEAR_BIT in the config mask.
This enables using the BILINEAR filter for scaling and
improves the image quality
Change-Id: Ic9b61fe3ed5b16e63971fe542928640a13d92764
CRs-fixed: 453925
When blending translucent layers or transitioning with c2d,
sometimes the frame is blended with stale framebuffer content.
Clear the frame with C2D fill surface before drawing.
CRs-fixed: 462461
Change-Id: I1dca477fe88d77f4f0fa4846b155404f348e911d
With the new C2D optimization, the layer transformation is
passed through the config_mask. If two layers had same
transform then the transform info was lost, fix that by
updating the config_mask when the layer has transformation info
Change-Id: Idecb5b16a65269fd13eead9a2de27a4756435583
CRs-fixed: 453083
- flush_get_fence API to copybit.h - which is async,
which returns the fenceFD
- flush_get_fence calls C2dflush and c2dCreateFenceFD
signals the c2d_wait_thread which waits for
transcation to finish and cleanup resources
Change-Id: I98d5e08ea1cbce9732970c68c1e47b6f396249ce
Single commit having all the following commits
commit f39e512ad6f51cf35ab00995e52a39f31e76e77f
Author: Naomi Luis <nluis@codeaurora.org>
Date: Fri Oct 19 14:41:32 2012 -0700
copybit_c2d: Use a single draw call for surfaces with different rotations
- When opening the copybit module, query the c2d driver capabilities.
- If supported, use a single draw call for blitting surfaces with different
orientations using the target target rotate override flag.
- If not supported, invoke finish_copybit to draw the layers with the
previous orientation before applying the current orientation.
Change-Id: I0446f2817786321c9caa2eabf7b6679e5ecbcabe
commit a4008ff75150c1591c0dca46354e6bb8b30a675b
Author: Naomi Luis <nluis@codeaurora.org>
Date: Thu Oct 18 15:26:15 2012 -0700
Display: Add copybit_c2d performance optimizations
Following performance optimizations have been made to copybit_c2d
- Club all the layers into a single draw call. When finish_copybit
is invoked, the draw and finish are executed.
- Send the layer blend mode to copybit.
- If the blend is set to none, set the pre-multiplied alpha flag.
SurfaceFlinger assumes all surfaces have pre-multiplied alpha unless
explicitly specified. Copybit_c2d does the same.
The following cases are currently exempt from this optimization:
- Change in C2D target
- The template surfaces limit has been reached
- Max. blit object limit has been reached.
- Change in target transform.
- Use of any temporary source/destination surfaces.
Change-Id: I5ad1dc48bbe6d38d319116ccaa800fefb4fbf9f6
commit b188d739f69272c4820e1a33c5e10aa2f2825a5f
Author: Naomi Luis <nluis@codeaurora.org>
Date: Mon Oct 15 11:17:11 2012 -0700
copybit: Add finish API
Add the finish API to copybit. This API is used to indicate the
end of the drawing required by the copybit client.
Change-Id: Ica4d13dbe6e7aeb7f88aabb734bf03e86c043acc
commit ecd56ace395fce78c14e2e753d3221218d062899
Author: Naomi Luis <nluis@codeaurora.org>
Date: Mon Oct 15 11:21:08 2012 -0700
copybit: Clean up copybit_c2d
- Remove unused functions
- Do nothing for the COPYBIT_BLIT_TO_FRAMEBUFFER case when using c2d
composition.
Change-Id: Idc416ce3742bb41f060e52ac70add44e584032b9
commit 35d45d0cff5f6fe866fcb5df71444824a45e046e
Author: Naomi Luis <nluis@codeaurora.org>
Date: Wed Oct 17 10:37:33 2012 -0700
copybit: Use correct alignment in get_size()
get_size() is used to get the size required by c2d. It's alignment
should be 32 and not 16.
Change-Id: I9fedca0186d0ea19883b967e127417b059621a99
commit 31311da328cbe2229fcf8d8d8e276ec64005bcb3
Author: Ramakant Singh <ramaka@codeaurora.org>
Date: Thu Sep 20 15:58:29 2012 +0530
HWC: use uncached buffer for tmp yv12 and tmp stretch buffer.
Change-Id: Iecc2a9c3c9a365d00d3491ce1007a36484209346
CRs-fixed: 396574
commit 5503c90b5001aa4d863139eebcd9c0c43f265702
Author: Pawan Kumar <pavaku@codeaurora.org>
Date: Mon Sep 3 09:14:13 2012 +0530
copybit: Fix crash when copy_image is called with Tile format
Add check for copy_image failure case.
CRs-Fixed: 393122
Change-Id: I5765120e87a135795d771c5d331275a66ade7e23
commit 48975bc53702ecba66ecaa5e5769036bb91f5cf2
Merge: feb1cd6 ab25cb4
Author: Linux Build Service Account <lnxbuild@localhost>
Date: Sat Sep 1 13:36:47 2012 -0700
Merge "Display ::Use proper ion heap id" into jb
commit ab25cb4818d9b87886678972f2861bacc93c580a
Author: Ramakant Singh <ramaka@codeaurora.org>
Date: Thu Aug 9 14:23:39 2012 +0530
Display ::Use proper ion heap id
Change the MM heap to CAMERA heap for 7627a in use case
1.Intermediate stretch of low resolution clips
2.Software conversion of Y12 format
Change-Id: Ib705d3245f601256179e2dedccbc707df85eb431
CRs-Fixed: 383115
commit 90bcaadf41451d89fe80f96961ffc0a07a333145
Author: Prabhanjan Kandula <pkandula@codeaurora.org>
Date: Mon Aug 6 12:50:12 2012 +0530
libcopybit: fix YUV buffer alignments as per gralloc
buffer alignment in copybit should be in sync with
gralloc allignments while allocating buffer.
CRs-fixed: 377051
Change-Id: Ib2ae64e368ea3c92d3494c71da605197ccb4a9a5
Conflicts:
libcopybit/copybit_c2d.cpp
commit 95cf0038ea6eb02314024d673418464f0739ddb4
Author: Sravan Kumar D.V.N <sravank1@codeaurora.org>
Date: Thu Aug 2 14:37:37 2012 +0530
copybit : Add support for NV12_ENCODABLE format
Change-Id: I7087dcad2238a3c35cc15c28d75b76a9e3ca6718
Change-Id: I748bffb4215d27f609beec209af0a5047858a314
- Update the display HAL from Code Aurora Forum
- Add updated overlay library
- Enable HWC with basic video going through overlay
- Cleanup some files
Change-Id: I65c687c51be458cee71213c79e03eeda962d9086
