ubuntu-omap:tiopenmax-domx.git
6 years agoPackage 2.1-2 debian/2.1-2
Hervé Fache [Tue, 2 Oct 2012 09:28:56 +0000 (11:28 +0200)]
Package 2.1-2

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoNeed pkg-config to build
Hervé Fache [Tue, 2 Oct 2012 09:27:39 +0000 (11:27 +0200)]
Need pkg-config to build

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoPackage 2.1-1 debian/2.1-1
Hervé Fache [Thu, 13 Sep 2012 15:37:11 +0000 (15:37 +0000)]
Package 2.1-1

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoUpdate dependencies: now needs DRM and OMAPDRM
Hervé Fache [Thu, 13 Sep 2012 16:01:45 +0000 (16:01 +0000)]
Update dependencies: now needs DRM and OMAPDRM

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoMerge remote-tracking branch 'origin/upstream' into HEAD
Hervé Fache [Thu, 13 Sep 2012 15:48:55 +0000 (15:48 +0000)]
Merge remote-tracking branch 'origin/upstream' into HEAD

6 years agoMerge branch 'master' of gitorious.org:gstreamer-omap/domx into HEAD upstream/2.1
Hervé Fache [Thu, 13 Sep 2012 15:45:23 +0000 (15:45 +0000)]
Merge branch 'master' of gitorious.org:gstreamer-omap/domx into HEAD

6 years agocamera: add DRM support to create buffer for DCC sending
Hervé Fache [Tue, 11 Sep 2012 22:37:42 +0000 (22:37 +0000)]
camera: add DRM support to create buffer for DCC sending

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoconfigure: detect DRM and define HAVE_DRM
Hervé Fache [Tue, 11 Sep 2012 19:33:44 +0000 (19:33 +0000)]
configure: detect DRM and define HAVE_DRM

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoDCC: change path to /usr/share/dcc_binaries for GLP
Hervé Fache [Fri, 7 Sep 2012 13:21:57 +0000 (15:21 +0200)]
DCC: change path to /usr/share/dcc_binaries for GLP

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoPackage 2.0.1-2 debian/2.0.1-2
Xavier Boudet [Mon, 30 Jul 2012 14:29:11 +0000 (16:29 +0200)]
Package 2.0.1-2

Signed-off-by: Xavier Boudet <x-boudet@ti.com>
6 years agoConfigure patch
Xavier Boudet [Mon, 30 Jul 2012 14:26:56 +0000 (16:26 +0200)]
Configure patch

Signed-off-by: Xavier Boudet <x-boudet@ti.com>
6 years agoPackage 2.0.1-1 debian/2.0.1-1
Xavier Boudet [Thu, 19 Jul 2012 14:48:46 +0000 (16:48 +0200)]
Package 2.0.1-1

Signed-off-by: Xavier Boudet <x-boudet@ti.com>
6 years agoUpdate debian files for new upstream delivery
Xavier Boudet [Thu, 19 Jul 2012 14:46:40 +0000 (16:46 +0200)]
Update debian files for new upstream delivery

Signed-off-by: Xavier Boudet <x-boudet@ti.com>
6 years agoMerge branch 'upstream'
Xavier Boudet [Thu, 19 Jul 2012 14:16:16 +0000 (16:16 +0200)]
Merge branch 'upstream'

6 years agoMerge remote-tracking branch 'upstream-tree/master' into upstream upstream/2.0.1
Xavier Boudet [Thu, 19 Jul 2012 14:12:43 +0000 (16:12 +0200)]
Merge remote-tracking branch 'upstream-tree/master' into upstream

Remote tree: git://gitorious.org/gstreamer-omap/domx.git
Commit: 0c706094f5d65a7e88e2b502957220a2a8f4fd31

6 years agoReset tree before git://gitorious.org/gstreamer-omap/domx.git import
Xavier Boudet [Thu, 19 Jul 2012 14:11:46 +0000 (16:11 +0200)]
Reset tree before git://gitorious.org/gstreamer-omap/domx.git import

Signed-off-by: Xavier Boudet <x-boudet@ti.com>
6 years agoAdded data type for HMS Gamma.
Yordan Lilov [Tue, 26 Jun 2012 13:46:47 +0000 (16:46 +0300)]
Added data type for HMS Gamma.

This data will be transferred with extra data
for each frame.
To request it, use OMX_TI_HMSGamma.

Change-Id: If0b1f68a267fd20901fa6072c52889770011fa65

6 years agoDOMX: Simultaneous sensors in Gesture Mode. This patch matches with dicati patch...
Yordan Lilov [Tue, 29 May 2012 07:49:00 +0000 (10:49 +0300)]
DOMX: Simultaneous sensors in Gesture Mode. This patch matches with dicati patch "Simultaneous RAW+YUV sensor for gesture" ( I9b5414d6ab41f091820ba40d5c7dabbf41563617 )

Change-Id: I388d2e0193d6f8b756fbc09ed8cce5dc53c93056

6 years agoAdd RAW face detection index.
Yordan Lilov [Thu, 17 May 2012 14:29:24 +0000 (17:29 +0300)]
Add RAW face detection index.

Synchronize extradata enum with remote ducati definitions
including raw face detection index.

Change-Id: I0847d605f53c7385b4fb59987cce8b007f628ec6
Signed-off-by: Yordan Lilov <ylilov@mm-sol.com>
Signed-off-by: Alberto Aguirre <a-aguirre@ti.com>
6 years agoDOMX - Added header for 3A export.
Yordan Lilov [Wed, 9 May 2012 08:53:47 +0000 (11:53 +0300)]
DOMX - Added header for 3A export.

Change-Id: I64b1bc7a72d59f50b61a223696e5022722a5b9a9
Signed-off-by: Yordan Lilov <ylilov@mm-sol.com>
6 years agoAdd pkgconfig support
Hervé Fache [Tue, 17 Jul 2012 09:43:18 +0000 (11:43 +0200)]
Add pkgconfig support

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoSet libs version to 0.1
Hervé Fache [Tue, 17 Jul 2012 10:43:45 +0000 (12:43 +0200)]
Set libs version to 0.1

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoRPC: add generic way to pass more than one buffer to free
Hervé Fache [Wed, 4 Jul 2012 14:57:24 +0000 (16:57 +0200)]
RPC: add generic way to pass more than one buffer to free

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoClean up a bit
Hervé Fache [Tue, 3 Jul 2012 18:03:45 +0000 (20:03 +0200)]
Clean up a bit

Signed-off-by: root <root@panda-harv.(none)>
6 years agoProxy: add support for DMAbuf
Hervé Fache [Mon, 2 Jul 2012 16:03:46 +0000 (18:03 +0200)]
Proxy: add support for DMAbuf

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoAdd missing include file
Hervé Fache [Tue, 26 Jun 2012 13:53:34 +0000 (15:53 +0200)]
Add missing include file

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoAdd missing #ifdef's
Hervé Fache [Tue, 26 Jun 2012 13:51:45 +0000 (15:51 +0200)]
Add missing #ifdef's

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoRevert "DOMX: framework to dynamically enable/disable perf trace + OMX buffer events...
Hervé Fache [Mon, 2 Jul 2012 16:05:47 +0000 (18:05 +0200)]
Revert "DOMX: framework to dynamically enable/disable perf trace + OMX buffer events trace"

This reverts commit fb9c61ef1f54f30eac80b62e443ce5c9efddd6df.

6 years ago[BUILD] solve the redefined __packed issue
Hervé Fache [Wed, 11 Jul 2012 08:58:19 +0000 (10:58 +0200)]
[BUILD] solve the redefined __packed issue

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years ago[BUILD] add include path for rpmsg_omx.h
Hervé Fache [Wed, 11 Jul 2012 08:20:00 +0000 (10:20 +0200)]
[BUILD] add include path for rpmsg_omx.h

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years ago[BUILD] add autotools support
Hervé Fache [Tue, 10 Jul 2012 16:37:15 +0000 (18:37 +0200)]
[BUILD] add autotools support

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years ago[BUILD] remove all makefiles
Hervé Fache [Tue, 10 Jul 2012 16:23:07 +0000 (18:23 +0200)]
[BUILD] remove all makefiles

Signed-off-by: Hervé Fache <h-fache@ti.com>
6 years agoDOMX: Adds HQ video operating mode
Emilian Peev [Tue, 5 Jun 2012 12:20:00 +0000 (15:20 +0300)]
DOMX: Adds HQ video operating mode

- Adds an additional entry for VideoHQ
  mode inside the OMX_CAMOPERATINGMODETYPE
  enum. This is a new use-case for OMAP5
  that utilizes the Bayer scaler.

Corresponding Ducati Patch:
[OMX Camera] Video HQ mode- Initial version
https://gerrit.ext.ti.com/gerrit/omap/6990

Change-Id: I67da64688f866a35f5e0768eb333511895b4134b
Signed-off-by: Emilian Peev <epeev@mm-sol.com>
6 years agoDOMX: Changes for secure playback ION handle mapping
Prasanna Kumar M.R [Thu, 28 Jun 2012 18:25:24 +0000 (13:25 -0500)]
DOMX: Changes for secure playback ION handle mapping

In secure playback case, we dont map the ion buffers. When we pass
the Ion handle, the Rpmsg fails in buffer registration. Hence
we need to share the ion buffer and then pass it to Rpmsg/Ducati.

Change-Id: I2bea9323aa557d340c2a2c2d653fd5f0480160a4
Signed-off-by: Prasanna Kumar M.R. <prasanna.kumarmr@ti.com>
Ported-by: Bryan Buckley <bryan.buckley@ti.com>
6 years agoWifi Display: Remove pipe write when no color conversion is done
Vidhoon Viswanathan [Mon, 18 Jun 2012 08:16:57 +0000 (01:16 -0700)]
Wifi Display: Remove pipe write when no color conversion is done

In case of Wifi Display usecase, where no color conversion is needed,
no buffers are created locally in DOMX. The pipe used for managing
locally created buffers was wrongly being written to after EBD, even
though there are no buffers to be tracked. This patch excludes the write
to pipe call in cases where no local buffers are created for color conv.

Change-Id: I527ecca7d7d979e15eae66cdddf069cf87cfd8db
Signed-off-by: Vidhoon Viswanathan <vidhoon@ti.com>
6 years ago[FIX] Check if buffer header exists before restoring buffer size
Vidhoon Viswanathan [Tue, 26 Jun 2012 06:38:58 +0000 (23:38 -0700)]
[FIX] Check if buffer header exists before restoring buffer size

In case of receiving invalid metadata, buffer sizes are restored
by the patch http://review.omapzoom.org/#/c/23320

This should be done only if the buffer header structure  is not NULL.
This check was missed by that patch and this patch fixes.

Change-Id: I0e0c896058f18866ec1c20225647f52f136a84a2
Signed-off-by: Vidhoon Viswanathan <vidhoon@ti.com>
6 years agoDOMX: framework to dynamically enable/disable perf trace + OMX buffer events trace
Frederic Turgis [Mon, 18 Jun 2012 12:17:06 +0000 (14:17 +0200)]
DOMX: framework to dynamically enable/disable perf trace + OMX buffer events trace

Added framework to control activation of perf trace:
- set DEBUG_DOMX_KPI_STATUS environment variable
- or set debug.domx.kpi_status android property. Env variable takes precedence

kpi_status is a bitfield to control enabling of various perf traces.
Currently only OMX buffers events are available with bitfield value of 1. It will
trace EmptyThisBuffer, FillThisBuffer, EmptyBufferDone, FillBufferDone.
It is aligned with Ducati equivalent trace. Enabling the 2 allows computing
MPU<->Ducati latency

Following parameters are included for each message:
- message type: ETB, FTB, EBD, FBD
- OMX component name: DECODE, H264E, CAMERA, ...
- number of such requests for this component
- time in us, from clock_gettime(CLOCK_MONOTONIC, ...)
- Remote buffer header to match Ducati trace
- FBD nTimeStamp metadata to identify an FBD buffer among framework SW layers

visual post-processing is described at
http://opbuwiki.dal.design.ti.com/index.php/Benelli-Ducati_performance_instrumentation
It is the same for MPU or Ducati trace

Example:
- setprop debug.domx.trace_level 3 (perf trace is trace level >= 3)
- setprop debug.domx.kpi_status 1
- logcat DOMX:D *:S

Change-Id: I3b8a996fb861ba84a9f96c98285a30209880423a
Signed-off-by: Frederic Turgis <f-turgis@ti.com>
6 years agoDOMX: Update trace level at component init and introduce PROFILING level
Frederic Turgis [Tue, 12 Jun 2012 17:39:36 +0000 (19:39 +0200)]
DOMX: Update trace level at component init and introduce PROFILING level

  * Profiling trace level introduced after WARN and before INFO levels
  * Cleaning of trace level usage through enum
  * "Trace level update" functionality called at every proxy common init
  * Trace level can be updated through Android property debug.domx.trace_level.
    Environment variable still has precedence over Android property

To modify trace level, do on terminal either:
- export TIMM_OSAL_TRACE_DEBUG_LEVEL=<trace level>
- setprop debug.domx.trace_level <trace level> (Android only)
For example setprop debug.domx.trace_level 3 to set PROFILING level

The new value is taken into account only at init of a new component.

Change-Id: I04de66a6160c11e9339a8f5a63ca0676755a17e0
Signed-off-by: Frederic Turgis <f-turgis@ti.com>
6 years ago[DOMX]: Fix error handling in RPC_sendPacket_sync
Vidhoon Viswanathan [Wed, 2 May 2012 11:59:13 +0000 (04:59 -0700)]
[DOMX]: Fix error handling in RPC_sendPacket_sync

Ducati faulty state is determined by the following
inference condition:
if(status<0 && errno==ENXIO)
{
//ducati in faulty state
}
After write call, errno was checked in wrong manner.
Also, write call returns number of bytes written.
This patch checks errno in case of write call failure
and detects faulty state of ducati. And it uses
status == nPacketSize as a tighter condition to validate
write call success.

Change-Id: I2062e128c4754b0faef2c33ca9465129105ade0c
Signed-off-by: Vidhoon Viswanathan <vidhoon@ti.com>
6 years agoH264 enc: Restore buffer size if invalid metadata received
Volodymyr Babchuk [Mon, 18 Jun 2012 09:28:15 +0000 (12:28 +0300)]
H264 enc: Restore buffer size if invalid metadata received

LOCAL_PROXY_H264E_EmptyThisBuffer() changes buffer size based on information
from DUCATI. If buffer contains metadata for encoder, buffer size is
inconsistent and wee need to change it back. This was not done if DUCATI passes
invalid metadata type.

This patch ensures, that in any case buffer sizes will be restored.

Change-Id: I17b00da3e5bb07d381f8c3bac07a5d5532a9b17d
Signed-off-by: Volodymyr Babchuk <volodymyr.babchuk@ti.com>
Signed-off-by: Volodymyr Frolov <volodymyr.frolov@ti.com>
6 years agoDOMX: Use Tiler-1D buffer when ION 1D allocation fails
Devaraj Rangasamy [Tue, 5 Jun 2012 12:45:26 +0000 (18:15 +0530)]
DOMX: Use Tiler-1D buffer when ION 1D allocation fails

By default Camera proxy and DOMX components uses ION 1D
heap for buffer allocation. For 512MB system configuration,
ION 1D heap maynot available. So use tiler 1D buffers as
fallback mechanism. Applicable for both 1GB and 512MB
configuration.

Change-Id: Iff9b264f13d03f99975c4ff9dce9d474edeac34c
Signed-off-by: Devaraj Rangasamy <dev@ti.com>
6 years agoenc proxy: only register EncoderMetadataPointers buffers
Tyler Luu [Fri, 8 Jun 2012 20:21:30 +0000 (15:21 -0500)]
enc proxy: only register EncoderMetadataPointers buffers

Only EncoderMetadataPointers need to be registered with rpc at
EmptyThisBuffer because only EncoderMetadataPointers are dynamically
changing the pBuffer to point to Gralloc pointers.

Change-Id: Iedcccc09cb361cba2cccaa06385b9e6885508a0a
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agocamera proxy: fix computation of component buffer size
Tyler Luu [Wed, 23 May 2012 23:57:04 +0000 (18:57 -0500)]
camera proxy: fix computation of component buffer size

we don't need to take format into consideration when calculating
allocation size for component buffers. ducati has already taken
care of this when sending us data for width and alloc lines.
we simply just need to compute the product.

Change-Id: I31d051e96e4571f02798a009651f32b34c066992
Signed-off-by: Emilian Peev <epeev@mm-sol.com>
6 years agodomx: sync interface
Tyler Luu [Wed, 23 May 2012 23:22:24 +0000 (18:22 -0500)]
domx: sync interface

commit: 90fa398 [OMAP-Camera-IQ] Disable Interp LSC for OV5650
branch: ducati_mmInc2.1

Change-Id: I9f87066ae35d5db191c10264bc6a3ca70b58c2fd
Signed-off-by: Vladimir Petrov <vppetrov@mm-sol.com>
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agoproxy common: make register/unregister conditional
Tyler Luu [Mon, 21 May 2012 19:52:21 +0000 (14:52 -0500)]
proxy common: make register/unregister conditional

since some kernels do not support OMX_IOCPVRREGISTER, we are
making the compilation on whether or not it is defined.

If OMX_IOCPVRREGISTER is defined in FS but return ENOTTY, this
means that kernel support for this new ioctl has not been merged
yet.

Change-Id: I691a0ba53a53c443efa5341dae93f429f60705d0
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agocamera proxy: pass ion fd's instead of handles
Tyler Luu [Wed, 16 May 2012 08:44:21 +0000 (03:44 -0500)]
camera proxy: pass ion fd's instead of handles

proxy common now requires ion fd's to be passed instead of
ion handles.

Change-Id: Ibc2c8b43ff1556643c533530f93143b9529947ac
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agoproxy: add support for sgx register ioctl
Tyler Luu [Tue, 22 May 2012 19:11:02 +0000 (14:11 -0500)]
proxy: add support for sgx register ioctl

rpmsg_omx requires clients to register gralloc (sgx) buffers
using seperate ioctl (OMX_IOCPVRREGISTER). Handles received
from regsiter need to be released using OMX_IOCIONUNREGISTER
when done.

the exploit of passing ion handles is no longer supported directly
to rpmsg is no longer supported. rpmsg mappings will fail if handles
are not first registered.

Change-Id: I8d2bad7a9a6a7c34a96bef494d4327e988c78e8b
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agoproxy common: choose buffer 1st in AllocateBuffer
Tyler Luu [Wed, 25 Apr 2012 22:03:54 +0000 (17:03 -0500)]
proxy common: choose buffer 1st in AllocateBuffer

Moving choose buffer logic to the beginning of
AllocateBuffer so we can use it to index tBufList
array when saving ion handle information

Fixes an issue missed by "domx: fix video record"

Change-Id: Icf5ddcda8b2ffdc67f5ac337b4e7b683b90c053f
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agodomx: pass registered buffer address to FreeBuffer
Tyler Luu [Fri, 13 Apr 2012 23:26:53 +0000 (18:26 -0500)]
domx: pass registered buffer address to FreeBuffer

Pass registered buffer address to RPC_FreeBuffer in
PROXY_FreeBuffer. rpmsg expects a registered address to
be passed to it when mapping is required.

Change-Id: I50abeaf5a7cfc31298832294648a0b8991cea3ad
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agodomx: fix video record
Tyler Luu [Fri, 30 Mar 2012 00:44:56 +0000 (19:44 -0500)]
domx: fix video record

With cpcam kernel changes for rpmsg driver, rpmsg omx
is requiring all buffers be registered and unregistered
starting here: http://review.omapzoom.org/#/c/20916

Since encoders are using this little trick of dynamically
reassigning handles for each input buffer, we will need
to register/unregister the handles every ETB.

Change-Id: Iaefcd8d3ba56605edfc398814bdec9a6524fa196
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agoMerge "domx: add register/unregister for buffers"
Keith Deacon [Thu, 24 May 2012 22:15:09 +0000 (17:15 -0500)]
Merge "domx: add register/unregister for buffers"

6 years agoMerge "domx: fixes for reprocessing"
Keith Deacon [Thu, 24 May 2012 22:14:13 +0000 (17:14 -0500)]
Merge "domx: fixes for reprocessing"

6 years agoMerge "domx: Handle Ducati requests for internal allocation"
Keith Deacon [Thu, 24 May 2012 22:13:27 +0000 (17:13 -0500)]
Merge "domx: Handle Ducati requests for internal allocation"

6 years ago[VDEC] Adds Index to enable/disable Decoder Metadata
Shivaraj Shetty [Tue, 8 May 2012 02:08:19 +0000 (07:38 +0530)]
[VDEC] Adds Index to enable/disable Decoder Metadata

Adds OMX_TI_IndexParamVideoEnableMetadata index to configure
the codec metadata. Also adds interface structures for SEI, VUI
and MBInfo Metadata interpretation.

By default SEI and VUI are enabled (only H264D and SVCD),
MBInfo is disabled.

Dependent Ducati Patch:
[VDEC] Full Codec Metadata Export to the Application
https://gerrit.ext.ti.com/gerrit/omap/#change,6451

Change-Id: I85e3a67410e4d23f0c696cb955654f323136df2f
Signed-off-by: Shivaraj Shetty <shettyshivaraj@ti.com>
6 years agodomx: add register/unregister for buffers
Tyler Luu [Wed, 11 Apr 2012 18:36:56 +0000 (13:36 -0500)]
domx: add register/unregister for buffers

need to register/unregister ion and gralloc handles with
rpmsg before sending. this includes shared buffers used
in get/set config and parameters.

Change-Id: I23e4c09349c9c4f87c06b65e270b320dba9f0372
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years agodomx: fixes for reprocessing
Tyler Luu [Tue, 27 Mar 2012 00:42:52 +0000 (19:42 -0500)]
domx: fixes for reprocessing

1. For OMX_TI_IndexUseNativeBuffers, properly set gralloc
   usage for false case. If client set this index to true,
   a subsequent set to false was not being handled causing
   domx to misintrepret handles.
2. For OMX_TI_IndexUseBufferDescriptor, RPC SetParameters
   needs to be called so Ducati also knows about the
   descriptor. This index was previously only being handled
   internally.

Change-Id: I928dda6d5ae9b76682c3b60dce4375a1accf09e4
Signed-off-by: Tyler Luu <tluu@ti.com>
6 years ago[DOMX] Fix missing condition for color conversion
Vidhoon Viswanathan [Wed, 16 May 2012 10:20:17 +0000 (03:20 -0700)]
[DOMX] Fix missing condition for color conversion

The conditional check to initiate color conversion was incomplete.
A portion of it was missed during rebase and dependency changes of
the patch: http://review.omapzoom.org/#/c/20188/

This patch adds the missing part to fix the condition.

Change-Id: I121004a0e47d28422ec40863b57682d2288c7d5e
Signed-off-by: Vidhoon Viswanathan <vidhoon@ti.com>
6 years agodomx: Handle Ducati requests for internal allocation
Tyler Luu [Fri, 30 Mar 2012 23:33:09 +0000 (18:33 -0500)]
domx: Handle Ducati requests for internal allocation

Indicate to ducati that we want to use dynamically mapped system heap
buffers, and process ducati requests for tiler area to map those buffers
into.

Change-Id: Ic713fc3c65acd3f478e9b6772af7486ff5a3be59
Signed-off-by: David Schleef <ds@ti.com>
6 years agoReplacing OMAP_ENHANCMENT flag with USE_ITTIAM_AAC flag
Barry Woodward [Wed, 18 Apr 2012 22:49:47 +0000 (17:49 -0500)]
Replacing OMAP_ENHANCMENT flag with USE_ITTIAM_AAC flag

The OMAP_ENHANCEMENT flag is not needed for TI code.
Also including these lines cause an error in builds where
the Ittiam binaries are not present.  Adding USE_ITTIAM_FLAG
so that this can be configured in the BoardConfig.mk file

Change-Id: Ia3fdefb442dc73280ec2386aff57ca439cdec679
Signed-off-by: Barry Woodward <a0270655@ti.com>
6 years agoMerge "Disable DOMX for omap3 only"
Jean Johnson [Wed, 2 May 2012 10:08:38 +0000 (05:08 -0500)]
Merge "Disable DOMX for omap3 only"

6 years agoDisable DOMX for omap3 only
Jean Johnson [Wed, 2 May 2012 05:19:05 +0000 (00:19 -0500)]
Disable DOMX for omap3 only

Ensure that DOMX is disableld for the OMAP3 program.  Other programs,
based on OMAP4 and OMAP5 use the domx program.  Previously, only the
OMAP4 program had used DOMX.

Change-Id: I798f1dd7dcde776790cfa9c605c9f96e28ba4065
Signed-off-by: Jean Johnson <jean-johnson@ti.com>
6 years agoWifi Display: Color convert Input buffers based on Pixel format
Vidhoon Viswanathan [Tue, 17 Apr 2012 12:08:26 +0000 (05:08 -0700)]
Wifi Display: Color convert Input buffers based on Pixel format

This patch determines the pixel format of buffers at
LOCAL_PROXY_H264E_EmptyThisBuffer and allocates local buffers for color
conversion only if the format is not NV12. Before, based on color format
of supplying port, local buffers used to be allocated at
LOCAL_PROXY_H264E_AllocateBuffer. But color conversion is dependent on
pixel format of buffer and not color format of supplying port.

Change-Id: I05c7098d515502ac3a5ab2097a5894f05139691a
Signed-off-by: Vidhoon Viswanathan <vidhoon@ti.com>
6 years agoH264 enc: Send valid pointer to NV12 buffer even if it has zero filled length.
Andriy Chepurnyy [Thu, 12 Apr 2012 08:56:46 +0000 (11:56 +0300)]
H264 enc: Send valid pointer to NV12 buffer even if it has zero filled length.

Even in case buffer filled len sent by client was 0, proxy should set correct
NV12 ptr preventing error in rpmsg on buffer address translation.

This pacth fix DR OMAPS00266438

Change-Id: I470ffa3c3e8b23f262c34819ac9ae025a160dad3
Signed-off-by: Andriy Chepurnyy <x0155536@ti.com>
6 years agoDOMX: Added OMX_TI_StereoVideo to OMX_CAMOPERATINGMODETYPE
Pavel Nedev [Thu, 5 Apr 2012 13:09:31 +0000 (16:09 +0300)]
DOMX: Added OMX_TI_StereoVideo to OMX_CAMOPERATINGMODETYPE

Added OMX_TI_StereoVideo to OMX_CAMOPERATINGMODETYPE enum in OMX_TI_IVCommon.h

Corresponding Ducati side patches:
2.0: https://gerrit.ext.ti.com/gerrit/omap/#change,5802
2.1: https://gerrit.ext.ti.com/gerrit/omap/#change,5801
master: https://gerrit.ext.ti.com/gerrit/omap/#change,5800

Change-Id: I5a312620061582453c95ae788aea72023699c52c
Signed-off-by: Pavel Nedev <pnedev@mm-sol.com>
6 years agoEnumerator declaration correction
Brijesh Nekkare [Tue, 3 Apr 2012 10:30:18 +0000 (16:00 +0530)]
Enumerator declaration correction

Enumerator declaration correction for OMX_TI_INDEXTYPE

Signed-off-by: Brijesh Nekkare <b-nekkare@ti.com>
6 years agoEncoder: Add a custom index to retrieve the component handle
Brijesh Nekkare [Tue, 3 Apr 2012 09:49:52 +0000 (15:19 +0530)]
Encoder: Add a custom index to retrieve the component handle

For slice based processing between camera and encoder, a tunnel must be set up
between camera and encoder. Therefore, the application requires access to the
component handle. However, the handle is hidden in lower layers of the framework.
Therefore, added a custom Index that applications can use to read the component handle.

Change-Id: I8dd24abe5d78bf188b76d48d88586cb923146679
Signed-off-by: Jorge E. Solano <jsolano@ti.com>
Signed-off-by: Anu Sundararajan <sanuradha@ti.com>
Conflicts:

omx_core/inc/OMX_TI_Index.h

Signed-off-by: Brijesh Nekkare <b-nekkare@ti.com>
6 years agoAdd index OMX_TI_IndexParamComponentExpectedSuspensionState
Abhishek Ranka [Tue, 3 Apr 2012 07:28:06 +0000 (00:28 -0700)]
Add index OMX_TI_IndexParamComponentExpectedSuspensionState

Add support to query using structure OMX_PARAM_SUSPENSIONTYPE
over the index OMX_TI_IndexParamComponentExpectedSuspensionState to
indicate the present or anticipated suspension status of OMX
camera if the IL Client tried to continue using OMX camera. If
OMX-camera is or anticipates it would be blocked on resources
if transitioned from states Loaded to Idle or beyond, it returns
the eType as OMX_Suspended and IL Clients may choose not to
continue using OMX-camera at present and try to use later on.
If the resources are available at present and it already has
or anticipates to get the resources, then the eType is returned
as OMX_NotSuspended.

Signed-off-by: Abhishek Ranka <asranka@ti.com>
6 years agoDOMX: Pass the correct buf address in PROXY_FreeBuffer
Pavel Nedev [Wed, 21 Mar 2012 10:23:03 +0000 (12:23 +0200)]
DOMX: Pass the correct buf address in PROXY_FreeBuffer

This fixes the address passed from PROXY_FreeBuffer to RPC_FreeBuffer to be the same
that is passed to RPC_UseBuffer from PROXY_UseBuffer.

Change-Id: I8581ec7a6abfafcb8ef11dd360a6114ba4d704d2
Signed-off-by: Pavel Nedev <pnedev@mm-sol.com>
6 years ago[DOMX] Fix Buffer Map type during free buffer
Pradeep Venkatasubbarao [Thu, 22 Mar 2012 13:17:18 +0000 (18:47 +0530)]
[DOMX] Fix Buffer Map type during free buffer

The buffer sent from RPC calls (UseBuffer, AllocateBuffer, FreeBuffer)
will get mapped to approriate space depending on the map type that is
set in the packet.

Even though a pBuffer was sent in the packet, the map type
(RPC_OMX_MAP_INFO_TYPE) was set to RPC_OMX_MAP_INFO_NONE.
This resulted in conversion not being done in rpmsg.

This patch sets the RPC_OMX_MAP_INFO_TYPE correctly in the
packet to RPC_OMX_MAP_INFO_ONE_BUF. With this rpmsg will
convert the pBuffer address to ducati side (tiler) address.

Change-Id: I722b618687e178817e2e8ffdb9670e7a5f54d5e9
Signed-off-by: Pradeep Venkatasubbarao <pradeepv@ti.com>
6 years agoAdded support for Ittiam AAC and BSAC decoders
Sergii Shcherbakov [Tue, 6 Mar 2012 15:33:20 +0000 (17:33 +0200)]
Added support for Ittiam AAC and BSAC decoders

Change-Id: I8e0cfc195630d0dfcc2ba6d3c644c8d984d5c454
Signed-off-by: Sergii Shcherbakov <x0155537@ti.com>
6 years agoSync between A9 DOMX and Ducati OMX interface
Brijesh Nekkare [Thu, 22 Mar 2012 09:20:20 +0000 (14:50 +0530)]
Sync between A9 DOMX and Ducati OMX interface

This syncup is needed due to the CP-CAM changes

Signed-off-by: Brijesh Nekkare <b-nekkare@ti.com>
6 years ago[OMAP4-DOMX]Patch to add new index to control feature to skip grey frames
Shivaraj Shetty [Fri, 9 Mar 2012 06:39:16 +0000 (12:09 +0530)]
[OMAP4-DOMX]Patch to add new index to control feature to skip grey frames

Signed-off-by: Shivaraj Shetty <shettyshivaraj@ti.com>
6 years ago[cp-cam] enqueue vector shot interface changes
Naveen Kumar Penda [Thu, 8 Mar 2012 11:46:42 +0000 (17:16 +0530)]
[cp-cam] enqueue vector shot interface changes

Change-Id: I416657d5e9fa6d07829458c7f17b7b921fb8267f
Signed-off-by: Naveen Kumar Penda <pnaveen@ti.com>
6 years agoAdd interfaces for OMX component allocation support
Naveen Kumar Penda [Mon, 5 Mar 2012 16:00:16 +0000 (21:30 +0530)]
Add interfaces for OMX component allocation support

Signed-off-by: naveen kumar penda <a0393443@ti.com>
This patch enables a OMX component to export its internal
buffer requirements (multiple) and a OMX component user to
query them, allocate them as per the buffer type and set the
buffers. This is expected to happen per port and during
regular OMX buffer life cycle (Loaded>Idle and Port Enable)

ducati side changes:
https://gerrit.ext.ti.com/gerrit/omap/#change,4950

Change-Id: I88d70364cc6f18a676c0224457ef03cb48595379

6 years agoExtend buffer types in OMX_TI_BUFFERTYPE
Naveen Kumar Penda [Mon, 5 Mar 2012 15:56:13 +0000 (21:26 +0530)]
Extend buffer types in OMX_TI_BUFFERTYPE

Signed-off-by: naveen kumar penda <a0393443@ti.com>
ducati side changes:
https://gerrit.ext.ti.com/gerrit/omap/#change,4949

Change-Id: Idf9c679b8d325f2d396895bb99baadcf9df19e31

6 years agoDynamic algo disable
Naveen Kumar Penda [Mon, 5 Mar 2012 15:50:32 +0000 (21:20 +0530)]
Dynamic algo disable

Signed-off-by: naveen kumar penda <a0393443@ti.com>
Dynamic algo disable (OMX interface)

1. Need ability to turn off (post-demosaicking) denoising - we definitely want to skip NSF2 and YUV denoising
2. Need ability to turn off sharpening
3. Need ability to set final tone mapping to gamma=1/2.2, or at least to read back the lookup table
4. Need ability to turn off non-linear color processing, i.e. turn off 3D-LUT color mapping

Ducati side changes:
https://gerrit.ext.ti.com/gerrit/omap/#change,4945

Change-Id: I055e74c9554b812b745c9d5fc5b9a2c622b70105

6 years agoOMX interface for Port Tap Point
Naveen Kumar Penda [Mon, 5 Mar 2012 15:47:37 +0000 (21:17 +0530)]
OMX interface for Port Tap Point

Signed-off-by: naveen kumar penda <a0393443@ti.com>
Ducati side changes:
https://gerrit.ext.ti.com/gerrit/omap/#change,4943

Change-Id: I69a99ddd247b291f542a40209c95f5c8eec92b39

6 years agoRe-Process mode in single preview
Naveen Kumar Penda [Mon, 5 Mar 2012 15:38:38 +0000 (21:08 +0530)]
Re-Process mode in single preview

Signed-off-by: naveen kumar penda <a0393443@ti.com>
Ducati side change:
https://gerrit.ext.ti.com/gerrit/omap/#change,4941

Change-Id: I2961e61f169e8ec0ef3cef3e280b5f59ece6d728

6 years agoNew CPCam Usecase interface change
Naveen Kumar Penda [Mon, 5 Mar 2012 15:33:31 +0000 (21:03 +0530)]
New CPCam Usecase interface change

Change-Id: I4291a9effd6061be28a45246b7bb22d75a712ccc
Signed-off-by: naveen kumar penda <a0393443@ti.com>
Ducati side changes:
https://gerrit.ext.ti.com/gerrit/omap/4939

6 years agoInterface changes for absolute exposure-gain bracketing
Naveen Kumar Penda [Mon, 5 Mar 2012 15:27:17 +0000 (20:57 +0530)]
Interface changes for absolute exposure-gain bracketing

corresponding Ducati patch to be merged:
https://gerrit.ext.ti.com/gerrit/omap/#change,4938

Change-Id: I2709776c18a0563a3e7694379c42977edac11ac5
Signed-off-by: Naveen Kumar Penda <pnaveen@ti.com>
6 years agoUpdating DOMX Core corresponding to OMX Core
Saurabh Bipin Chandra [Fri, 2 Mar 2012 08:59:44 +0000 (14:29 +0530)]
Updating DOMX Core corresponding to OMX Core

Abstracting the dynamic Rate control params exposed at the OMX level by the Ducatimm patch
Change-Id: I5253e6db54115dbade0619ba274c39b5e461b814
to the A9 side.
Also adding a few indexes from the OMX core which were
missing in the DOMX core.

Signed-off-by: Saurabh Bipin Chandra <a0131926@ti.com>
6 years ago[domx-cam]: Buffer allocation for 1080p was smaller for VTC usecase
Phanish Hanagal [Thu, 19 Jan 2012 15:09:12 +0000 (20:39 +0530)]
[domx-cam]: Buffer allocation for 1080p was smaller for VTC usecase

Description: Buffer allocation for VTC usecase with VNF at 1080p was not
sufficient and hence the 1080p usecase was failing. This patch fixes the
allocation issue.

Change-Id: I8622738269e3bb92600669dc93f62f5ef3d820e8
Signed-off-by: Phanish Hanagal <phanish.hs@ti.com>
6 years agoDOMX: OMX param for sensor compression
Emilian Peev [Fri, 3 Feb 2012 09:39:52 +0000 (11:39 +0200)]
DOMX: OMX param for sensor compression

- Corresponding Ducati change:
  https://gerrit.ext.ti.com/gerrit/omap/#change,4346

Change-Id: I54e0ab344a57c214d1c5c1a94f5ff2e1bb6d5acb
Signed-off-by: Emilian Peev <epeev@mm-sol.com>
6 years agodebian/changelog debian/1.5.3.1-4
Olivier Naudan [Mon, 30 Jan 2012 14:47:27 +0000 (09:47 -0500)]
debian/changelog

6 years agoAdded two debian/patches, to fix a memory leak and push up a limit.
Olivier Naudan [Wed, 18 Jan 2012 15:25:37 +0000 (16:25 +0100)]
Added two debian/patches, to fix a memory leak and push up a limit.

6 years agoSetupTunnel:fix to return Comp returned Err value
Radha Dasari [Fri, 13 Jan 2012 06:51:11 +0000 (12:21 +0530)]
SetupTunnel:fix to return Comp returned Err value

corrected the structure value to be update : so that the
Comp returned Value will be notified to the APP

Signed-off-by: Radha Dasari <radhapurnima@ti.com>
6 years agoDOMX: Add new camera views
Ivan Evlogiev [Wed, 21 Dec 2011 13:29:01 +0000 (15:29 +0200)]
DOMX: Add new camera views

Depends on
ducati master
https://gerrit.ext.ti.com/gerrit/omap/#change,3310
ducati 2.0
https://gerrit.ext.ti.com/gerrit/omap/#change,3312

Change-Id: I2228a97ac75b5cce056e588fd8237870dda36588
Signed-off-by: Ivan Evlogiev <x0099089@ti.com>
6 years agoDOMX: Change use of structure HISTOGRAMTYPE
Ivan Evlogiev [Mon, 19 Dec 2011 14:47:44 +0000 (16:47 +0200)]
DOMX: Change use of structure HISTOGRAMTYPE

Till now OMX_HISTOGRAMTYPE and OMX_PARAM_SHAREDBUFFER
were used along/mixed with OMX_TI_HISTOGRAMTYPE and
OMX_TI_CONFIG_SHAREDBUFFER.
The fix set OMX_TI_HISTOGRAMTYPE and OMX_TI_CONFIG_SHAREDBUFFER
to be used only.
On A9 side only the HYSTOGRAMTYPE was not cleaned up

Ducati side patches are:
maser branch:
https://gerrit.ext.ti.com/gerrit/omap/#change,3248
ducati_mmInc2.0 branch:
https://gerrit.ext.ti.com/gerrit/omap/#change,3251

Change-Id: Ib9a9450a1f12702f4d27a4f55a386918f80a04ea
Signed-off-by: Ivan Evlogiev <x0099089@ti.com>
6 years ago[Code Clean] domx code cleanup
a kasi viswanathan [Mon, 19 Dec 2011 10:45:47 +0000 (16:15 +0530)]
[Code Clean] domx code cleanup

code clean up for domx, removing dead code
Reverting back omx_rpc_stub.c, timm_osal_pipes.c and
omx_proxy_common.c for future use and reference

Signed-off-by: a kasi viswanathan <kasi.alageswaran@ti.com>
6 years ago[TRACE] Include function name in tracing
a kasi viswanathan [Tue, 6 Dec 2011 07:56:36 +0000 (13:26 +0530)]
[TRACE] Include function name in tracing

Print path to function, line number and function name for trace levels 1, 2 and 5. (error, warning and entry/exit)
Modified to remove tabs and more spaces

Signed-off-by: a kasi viswanathan <kasi.alageswaran@ti.com>
6 years ago[Compiler warning removal]
a kasi viswanathan [Wed, 7 Dec 2011 05:55:18 +0000 (11:25 +0530)]
[Compiler warning removal]

Removal of compiler warnings.
typecastings of datatypes, adding default statements for switches, etc.

Signed-off-by: a kasi viswanathan <kasi.alageswaran@ti.com>
6 years agoMerge "[Cleanup] unused roles , unused components"
Pradeep Venkatasubbarao [Tue, 20 Dec 2011 11:30:02 +0000 (05:30 -0600)]
Merge "[Cleanup] unused roles , unused components"

6 years agoDOMX: Adds GLBCE to OMX Capabilities
Emilian Peev [Mon, 12 Dec 2011 09:25:56 +0000 (11:25 +0200)]
DOMX: Adds GLBCE to OMX Capabilities

Change-Id: I196fae46282778240d68331e0a4d1abefb8a8efc
Signed-off-by: Emilian Peev <epeev@mm-sol.com>
6 years ago[omx-venc]interface update -auto frame rate update
Radha Dasari [Fri, 9 Dec 2011 13:16:35 +0000 (18:46 +0530)]
[omx-venc]interface update -auto frame rate update

interface for the configuration of ato frame rate update
to the codec from component has been added.

OMX when client enbles it, the OMX component calculates
the change in frame rate & inform the codec accordingly.

the corresponding ducati patch is
https://gerrit.ext.ti.com/gerrit/omap/#change,3030

Signed-off-by: Radha Dasari <radhapurnima@ti.com>
6 years ago[DOMX] Adding check for null buffer at PROXY_FreeBuffer
Saurabh Bipin Chandra [Fri, 9 Dec 2011 09:55:39 +0000 (15:25 +0530)]
[DOMX] Adding check for null buffer at PROXY_FreeBuffer

Memory unmapping to be avoided for null buffers

Change-Id: I9c2a7fd1029f314f4feddaded981799de0a14a84
Signed-off-by: Saurabh Bipin Chandra <a0131926@ti.com>
6 years ago[Cleanup] unused roles , unused components
a kasi viswanathan [Tue, 6 Dec 2011 13:09:48 +0000 (18:39 +0530)]
[Cleanup] unused roles , unused components

updation of component table for removal of unused roles of components,
unused components and addition of new roles for components.

Signed-off-by: a kasi viswanathan <kasi.alageswaran@ti.com>
6 years agodebian/changelog debian/1.5.3.1-1
Olivier Naudan [Thu, 20 Oct 2011 10:58:31 +0000 (06:58 -0400)]
debian/changelog

6 years agoCFLAGS: do not treat compiler warnings as errors
Olivier Naudan [Fri, 14 Oct 2011 15:47:23 +0000 (11:47 -0400)]
CFLAGS: do not treat compiler warnings as errors

6 years agoPatch to fix segmentation fault in getconfig for SVC decoder
Shivaraj Shetty [Fri, 2 Dec 2011 05:08:24 +0000 (10:38 +0530)]
Patch to fix segmentation fault in getconfig for SVC decoder

This patch fixes the segmentation fault in DOMX while calling getconfig to query layer details for SVC decoder

Signed-off-by: Shivaraj Shetty <shettyshivaraj@ti.com>
6 years agoMerge "DOMX: fix an ION handle leak"
Sujatha Ratnala [Thu, 1 Dec 2011 12:21:03 +0000 (06:21 -0600)]
Merge "DOMX: fix an ION handle leak"