blob: baef9d2b9a4003944a8eda7b8ddad50f80eaab73 [file] [log] [blame]
/*
* Copyright (C) 2012-2013, The Linux Foundation. All rights reserved.
* Not a Contribution, Apache license notifications and license are retained
* for attribution purposes only.
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#include <math.h>
#include "hwc_mdpcomp.h"
#include <sys/ioctl.h>
#include "external.h"
#include "virtual.h"
#include "qdMetaData.h"
#include "mdp_version.h"
#include "hwc_fbupdate.h"
#include "hwc_ad.h"
#include <overlayRotator.h>
using namespace overlay;
using namespace qdutils;
using namespace overlay::utils;
namespace ovutils = overlay::utils;
namespace qhwc {
//==============MDPComp========================================================
IdleInvalidator *MDPComp::idleInvalidator = NULL;
bool MDPComp::sIdleFallBack = false;
bool MDPComp::sDebugLogs = false;
bool MDPComp::sEnabled = false;
bool MDPComp::sEnableMixedMode = true;
bool MDPComp::sEnablePartialFrameUpdate = false;
int MDPComp::sMaxPipesPerMixer = MAX_PIPES_PER_MIXER;
double MDPComp::sMaxBw = 0.0;
double MDPComp::sBwClaimed = 0.0;
bool MDPComp::sEnable4k2kYUVSplit = false;
MDPComp* MDPComp::getObject(hwc_context_t *ctx, const int& dpy) {
if(isDisplaySplit(ctx, dpy)) {
return new MDPCompSplit(dpy);
}
return new MDPCompNonSplit(dpy);
}
MDPComp::MDPComp(int dpy):mDpy(dpy){};
void MDPComp::dump(android::String8& buf)
{
if(mCurrentFrame.layerCount > MAX_NUM_APP_LAYERS)
return;
dumpsys_log(buf,"HWC Map for Dpy: %s \n",
(mDpy == 0) ? "\"PRIMARY\"" :
(mDpy == 1) ? "\"EXTERNAL\"" : "\"VIRTUAL\"");
dumpsys_log(buf,"CURR_FRAME: layerCount:%2d mdpCount:%2d "
"fbCount:%2d \n", mCurrentFrame.layerCount,
mCurrentFrame.mdpCount, mCurrentFrame.fbCount);
dumpsys_log(buf,"needsFBRedraw:%3s pipesUsed:%2d MaxPipesPerMixer: %d \n",
(mCurrentFrame.needsRedraw? "YES" : "NO"),
mCurrentFrame.mdpCount, sMaxPipesPerMixer);
dumpsys_log(buf," --------------------------------------------- \n");
dumpsys_log(buf," listIdx | cached? | mdpIndex | comptype | Z \n");
dumpsys_log(buf," --------------------------------------------- \n");
for(int index = 0; index < mCurrentFrame.layerCount; index++ )
dumpsys_log(buf," %7d | %7s | %8d | %9s | %2d \n",
index,
(mCurrentFrame.isFBComposed[index] ? "YES" : "NO"),
mCurrentFrame.layerToMDP[index],
(mCurrentFrame.isFBComposed[index] ?
(mCurrentFrame.drop[index] ? "DROP" :
(mCurrentFrame.needsRedraw ? "GLES" : "CACHE")) : "MDP"),
(mCurrentFrame.isFBComposed[index] ? mCurrentFrame.fbZ :
mCurrentFrame.mdpToLayer[mCurrentFrame.layerToMDP[index]].pipeInfo->zOrder));
dumpsys_log(buf,"\n");
}
bool MDPComp::init(hwc_context_t *ctx) {
if(!ctx) {
ALOGE("%s: Invalid hwc context!!",__FUNCTION__);
return false;
}
char property[PROPERTY_VALUE_MAX];
sEnabled = false;
if((property_get("persist.hwc.mdpcomp.enable", property, NULL) > 0) &&
(!strncmp(property, "1", PROPERTY_VALUE_MAX ) ||
(!strncasecmp(property,"true", PROPERTY_VALUE_MAX )))) {
sEnabled = true;
}
sEnableMixedMode = true;
if((property_get("debug.mdpcomp.mixedmode.disable", property, NULL) > 0) &&
(!strncmp(property, "1", PROPERTY_VALUE_MAX ) ||
(!strncasecmp(property,"true", PROPERTY_VALUE_MAX )))) {
sEnableMixedMode = false;
}
if(property_get("debug.mdpcomp.logs", property, NULL) > 0) {
if(atoi(property) != 0)
sDebugLogs = true;
}
if(property_get("persist.hwc.partialupdate.enable", property, NULL) > 0) {
if((atoi(property) != 0) && ctx->mMDP.panel == MIPI_CMD_PANEL &&
qdutils::MDPVersion::getInstance().is8x74v2())
sEnablePartialFrameUpdate = true;
}
ALOGE_IF(isDebug(), "%s: Partial Update applicable?: %d",__FUNCTION__,
sEnablePartialFrameUpdate);
sMaxPipesPerMixer = MAX_PIPES_PER_MIXER;
if(property_get("debug.mdpcomp.maxpermixer", property, "-1") > 0) {
int val = atoi(property);
if(val >= 0)
sMaxPipesPerMixer = min(val, MAX_PIPES_PER_MIXER);
}
if(ctx->mMDP.panel != MIPI_CMD_PANEL) {
// Idle invalidation is not necessary on command mode panels
long idle_timeout = DEFAULT_IDLE_TIME;
if(property_get("debug.mdpcomp.idletime", property, NULL) > 0) {
if(atoi(property) != 0)
idle_timeout = atoi(property);
}
//create Idle Invalidator only when not disabled through property
if(idle_timeout != -1)
idleInvalidator = IdleInvalidator::getInstance();
if(idleInvalidator == NULL) {
ALOGE("%s: failed to instantiate idleInvalidator object",
__FUNCTION__);
} else {
idleInvalidator->init(timeout_handler, ctx, idle_timeout);
}
}
if((property_get("debug.mdpcomp.4k2kSplit", property, "0") > 0) &&
(!strncmp(property, "1", PROPERTY_VALUE_MAX ) ||
(!strncasecmp(property,"true", PROPERTY_VALUE_MAX )))) {
sEnable4k2kYUVSplit = true;
}
return true;
}
void MDPComp::reset(const int& numLayers, hwc_display_contents_1_t* list) {
mCurrentFrame.reset(numLayers);
mCachedFrame.cacheAll(list);
mCachedFrame.updateCounts(mCurrentFrame);
}
void MDPComp::timeout_handler(void *udata) {
struct hwc_context_t* ctx = (struct hwc_context_t*)(udata);
if(!ctx) {
ALOGE("%s: received empty data in timer callback", __FUNCTION__);
return;
}
if(!ctx->proc) {
ALOGE("%s: HWC proc not registered", __FUNCTION__);
return;
}
sIdleFallBack = true;
/* Trigger SF to redraw the current frame */
ctx->proc->invalidate(ctx->proc);
}
void MDPComp::setMDPCompLayerFlags(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
LayerProp *layerProp = ctx->layerProp[mDpy];
for(int index = 0; index < ctx->listStats[mDpy].numAppLayers; index++) {
hwc_layer_1_t* layer = &(list->hwLayers[index]);
if(!mCurrentFrame.isFBComposed[index]) {
layerProp[index].mFlags |= HWC_MDPCOMP;
layer->compositionType = HWC_OVERLAY;
layer->hints |= HWC_HINT_CLEAR_FB;
} else {
/* Drop the layer when its already present in FB OR when it lies
* outside frame's ROI */
if(!mCurrentFrame.needsRedraw || mCurrentFrame.drop[index]) {
layer->compositionType = HWC_OVERLAY;
}
}
}
}
MDPComp::FrameInfo::FrameInfo() {
reset(0);
}
void MDPComp::FrameInfo::reset(const int& numLayers) {
for(int i = 0 ; i < MAX_PIPES_PER_MIXER && numLayers; i++ ) {
if(mdpToLayer[i].pipeInfo) {
delete mdpToLayer[i].pipeInfo;
mdpToLayer[i].pipeInfo = NULL;
//We dont own the rotator
mdpToLayer[i].rot = NULL;
}
}
memset(&mdpToLayer, 0, sizeof(mdpToLayer));
memset(&layerToMDP, -1, sizeof(layerToMDP));
memset(&isFBComposed, 1, sizeof(isFBComposed));
layerCount = numLayers;
fbCount = numLayers;
mdpCount = 0;
needsRedraw = true;
fbZ = 0;
}
void MDPComp::FrameInfo::map() {
// populate layer and MDP maps
int mdpIdx = 0;
for(int idx = 0; idx < layerCount; idx++) {
if(!isFBComposed[idx]) {
mdpToLayer[mdpIdx].listIndex = idx;
layerToMDP[idx] = mdpIdx++;
}
}
}
MDPComp::LayerCache::LayerCache() {
reset();
}
void MDPComp::LayerCache::reset() {
memset(&hnd, 0, sizeof(hnd));
memset(&isFBComposed, true, sizeof(isFBComposed));
memset(&drop, false, sizeof(drop));
layerCount = 0;
}
void MDPComp::LayerCache::cacheAll(hwc_display_contents_1_t* list) {
const int numAppLayers = list->numHwLayers - 1;
for(int i = 0; i < numAppLayers; i++) {
hnd[i] = list->hwLayers[i].handle;
}
}
void MDPComp::LayerCache::updateCounts(const FrameInfo& curFrame) {
layerCount = curFrame.layerCount;
memcpy(&isFBComposed, &curFrame.isFBComposed, sizeof(isFBComposed));
memcpy(&drop, &curFrame.drop, sizeof(drop));
}
bool MDPComp::LayerCache::isSameFrame(const FrameInfo& curFrame,
hwc_display_contents_1_t* list) {
if(layerCount != curFrame.layerCount)
return false;
for(int i = 0; i < curFrame.layerCount; i++) {
if((curFrame.isFBComposed[i] != isFBComposed[i]) ||
(curFrame.drop[i] != drop[i])) {
return false;
}
if(curFrame.isFBComposed[i] &&
(hnd[i] != list->hwLayers[i].handle)){
return false;
}
}
return true;
}
bool MDPComp::isSupportedForMDPComp(hwc_context_t *ctx, hwc_layer_1_t* layer) {
private_handle_t *hnd = (private_handle_t *)layer->handle;
if((not isYuvBuffer(hnd) and has90Transform(layer)) or
(not isValidDimension(ctx,layer))
//More conditions here, SKIP, sRGB+Blend etc
) {
return false;
}
return true;
}
bool MDPComp::isValidDimension(hwc_context_t *ctx, hwc_layer_1_t *layer) {
const int dpy = HWC_DISPLAY_PRIMARY;
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(!hnd) {
if (layer->flags & HWC_COLOR_FILL) {
// Color layer
return true;
}
ALOGE("%s: layer handle is NULL", __FUNCTION__);
return false;
}
//XXX: Investigate doing this with pixel phase on MDSS
if(!isSecureBuffer(hnd) && isNonIntegralSourceCrop(layer->sourceCropf))
return false;
int hw_w = ctx->dpyAttr[mDpy].xres;
int hw_h = ctx->dpyAttr[mDpy].yres;
hwc_rect_t crop = integerizeSourceCrop(layer->sourceCropf);
hwc_rect_t dst = layer->displayFrame;
int crop_w = crop.right - crop.left;
int crop_h = crop.bottom - crop.top;
int dst_w = dst.right - dst.left;
int dst_h = dst.bottom - dst.top;
float w_dscale = ceilf((float)crop_w / (float)dst_w);
float h_dscale = ceilf((float)crop_h / (float)dst_h);
/* Workaround for MDP HW limitation in DSI command mode panels where
* FPS will not go beyond 30 if buffers on RGB pipes are of width or height
* less than 5 pixels
* There also is a HW limilation in MDP, minimum block size is 2x2
* Fallback to GPU if height is less than 2.
*/
if((crop_w < 5)||(crop_h < 5))
return false;
if((w_dscale > 1.0f) || (h_dscale > 1.0f)) {
const uint32_t downscale =
qdutils::MDPVersion::getInstance().getMaxMDPDownscale();
if(ctx->mMDP.version >= qdutils::MDSS_V5) {
/* Workaround for downscales larger than 4x.
* Will be removed once decimator block is enabled for MDSS
*/
if(!qdutils::MDPVersion::getInstance().supportsDecimation()) {
if(crop_w > MAX_DISPLAY_DIM || w_dscale > downscale ||
h_dscale > downscale)
return false;
} else {
if(w_dscale > 64 || h_dscale > 64)
return false;
}
} else { //A-family
if(w_dscale > downscale || h_dscale > downscale)
return false;
}
}
return true;
}
ovutils::eDest MDPComp::getMdpPipe(hwc_context_t *ctx, ePipeType type,
int mixer) {
overlay::Overlay& ov = *ctx->mOverlay;
ovutils::eDest mdp_pipe = ovutils::OV_INVALID;
switch(type) {
case MDPCOMP_OV_DMA:
mdp_pipe = ov.nextPipe(ovutils::OV_MDP_PIPE_DMA, mDpy, mixer);
if(mdp_pipe != ovutils::OV_INVALID) {
return mdp_pipe;
}
case MDPCOMP_OV_ANY:
case MDPCOMP_OV_RGB:
mdp_pipe = ov.nextPipe(ovutils::OV_MDP_PIPE_RGB, mDpy, mixer);
if(mdp_pipe != ovutils::OV_INVALID) {
return mdp_pipe;
}
if(type == MDPCOMP_OV_RGB) {
//Requested only for RGB pipe
break;
}
case MDPCOMP_OV_VG:
return ov.nextPipe(ovutils::OV_MDP_PIPE_VG, mDpy, mixer);
default:
ALOGE("%s: Invalid pipe type",__FUNCTION__);
return ovutils::OV_INVALID;
};
return ovutils::OV_INVALID;
}
bool MDPComp::isFrameDoable(hwc_context_t *ctx) {
bool ret = true;
const int numAppLayers = ctx->listStats[mDpy].numAppLayers;
if(!isEnabled()) {
ALOGD_IF(isDebug(),"%s: MDP Comp. not enabled.", __FUNCTION__);
ret = false;
} else if(qdutils::MDPVersion::getInstance().is8x26() &&
ctx->mVideoTransFlag && ctx->mVirtualDisplay->isConnected()) {
//1 Padding round to shift pipes across mixers
ALOGD_IF(isDebug(),"%s: MDP Comp. video transition padding round",
__FUNCTION__);
ret = false;
} else if(ctx->dpyAttr[HWC_DISPLAY_EXTERNAL].isConfiguring ||
ctx->dpyAttr[HWC_DISPLAY_VIRTUAL].isConfiguring) {
ALOGD_IF( isDebug(),"%s: External Display connection is pending",
__FUNCTION__);
ret = false;
} else if(ctx->isPaddingRound) {
ctx->isPaddingRound = false;
ALOGD_IF(isDebug(), "%s: padding round",__FUNCTION__);
ret = false;
}
return ret;
}
/*
* 1) Identify layers that are not visible in the updating ROI and drop them
* from composition.
* 2) If we have a scaling layers which needs cropping against generated ROI.
* Reset ROI to full resolution.
*/
bool MDPComp::validateAndApplyROI(hwc_context_t *ctx,
hwc_display_contents_1_t* list, hwc_rect_t roi) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
if(!isValidRect(roi))
return false;
hwc_rect_t visibleRect = roi;
for(int i = numAppLayers - 1; i >= 0; i--){
if(!isValidRect(visibleRect)) {
mCurrentFrame.drop[i] = true;
mCurrentFrame.dropCount++;
}
const hwc_layer_1_t* layer = &list->hwLayers[i];
hwc_rect_t dstRect = layer->displayFrame;
hwc_rect_t srcRect = integerizeSourceCrop(layer->sourceCropf);
int transform = layer->transform;
hwc_rect_t res = getIntersection(visibleRect, dstRect);
int res_w = res.right - res.left;
int res_h = res.bottom - res.top;
int dst_w = dstRect.right - dstRect.left;
int dst_h = dstRect.bottom - dstRect.top;
if(!isValidRect(res)) {
mCurrentFrame.drop[i] = true;
mCurrentFrame.dropCount++;
}else {
/* Reset frame ROI when any layer which needs scaling also needs ROI
* cropping */
if((res_w != dst_w || res_h != dst_h) &&
needsScaling (layer)) {
ALOGI("%s: Resetting ROI due to scaling", __FUNCTION__);
memset(&mCurrentFrame.drop, 0, sizeof(mCurrentFrame.drop));
mCurrentFrame.dropCount = 0;
return false;
}
}
if (layer->blending == HWC_BLENDING_NONE)
visibleRect = deductRect(visibleRect, res);
}
return true;
}
void MDPComp::generateROI(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
if(!sEnablePartialFrameUpdate) {
return;
}
if(mDpy || isDisplaySplit(ctx, mDpy)){
ALOGE_IF(isDebug(), "%s: ROI not supported for"
"the (1) external / virtual display's (2) dual DSI displays",
__FUNCTION__);
return;
}
if(isSkipPresent(ctx, mDpy))
return;
if(list->flags & HWC_GEOMETRY_CHANGED)
return;
struct hwc_rect roi = (struct hwc_rect){0, 0, 0, 0};
for(int index = 0; index < numAppLayers; index++ ) {
if ((mCachedFrame.hnd[index] != list->hwLayers[index].handle) ||
isYuvBuffer((private_handle_t *)list->hwLayers[index].handle)) {
hwc_rect_t dstRect = list->hwLayers[index].displayFrame;
hwc_rect_t srcRect = integerizeSourceCrop(
list->hwLayers[index].sourceCropf);
int transform = list->hwLayers[index].transform;
/* Intersect against display boundaries */
roi = getUnion(roi, dstRect);
}
}
if(!validateAndApplyROI(ctx, list, roi)){
roi = (struct hwc_rect) {0, 0,
(int)ctx->dpyAttr[mDpy].xres, (int)ctx->dpyAttr[mDpy].yres};
}
ctx->listStats[mDpy].roi.x = roi.left;
ctx->listStats[mDpy].roi.y = roi.top;
ctx->listStats[mDpy].roi.w = roi.right - roi.left;
ctx->listStats[mDpy].roi.h = roi.bottom - roi.top;
ALOGD_IF(isDebug(),"%s: generated ROI: [%d, %d, %d, %d]", __FUNCTION__,
roi.left, roi.top, roi.right, roi.bottom);
}
/* Checks for conditions where all the layers marked for MDP comp cannot be
* bypassed. On such conditions we try to bypass atleast YUV layers */
bool MDPComp::isFullFrameDoable(hwc_context_t *ctx,
hwc_display_contents_1_t* list){
const int numAppLayers = ctx->listStats[mDpy].numAppLayers;
if(sIdleFallBack && !ctx->listStats[mDpy].secureUI) {
ALOGD_IF(isDebug(), "%s: Idle fallback dpy %d",__FUNCTION__, mDpy);
return false;
}
if(isSkipPresent(ctx, mDpy)) {
ALOGD_IF(isDebug(),"%s: SKIP present: %d",
__FUNCTION__,
isSkipPresent(ctx, mDpy));
return false;
}
for(int i = 0; i < numAppLayers; ++i) {
hwc_layer_1_t* layer = &list->hwLayers[i];
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(isYuvBuffer(hnd) && has90Transform(layer)) {
if(!canUseRotator(ctx, mDpy)) {
ALOGD_IF(isDebug(), "%s: Can't use rotator for dpy %d",
__FUNCTION__, mDpy);
return false;
}
}
//For 8x26 with panel width>1k, if RGB layer needs HFLIP fail mdp comp
// may not need it if Gfx pre-rotation can handle all flips & rotations
if(qdutils::MDPVersion::getInstance().is8x26() &&
(ctx->dpyAttr[mDpy].xres > 1024) &&
(layer->transform & HWC_TRANSFORM_FLIP_H) &&
(!isYuvBuffer(hnd)))
return false;
}
if(ctx->mAD->isDoable()) {
return false;
}
//If all above hard conditions are met we can do full or partial MDP comp.
bool ret = false;
if(fullMDPComp(ctx, list)) {
ret = true;
} else if(partialMDPComp(ctx, list)) {
ret = true;
}
if(!hwLimitationsCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: HW limitations",__FUNCTION__);
return false;
}
return ret;
}
bool MDPComp::fullMDPComp(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
//Will benefit presentation / secondary-only layer.
if((mDpy > HWC_DISPLAY_PRIMARY) &&
(list->numHwLayers - 1) > MAX_SEC_LAYERS) {
ALOGD_IF(isDebug(), "%s: Exceeds max secondary pipes",__FUNCTION__);
return false;
}
const int numAppLayers = ctx->listStats[mDpy].numAppLayers;
for(int i = 0; i < numAppLayers; i++) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if(not isSupportedForMDPComp(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: Unsupported layer in list",__FUNCTION__);
return false;
}
//For 8x26, if there is only one layer which needs scale for secondary
//while no scale for primary display, DMA pipe is occupied by primary.
//If need to fall back to GLES composition, virtual display lacks DMA
//pipe and error is reported.
if(qdutils::MDPVersion::getInstance().is8x26() &&
mDpy >= HWC_DISPLAY_EXTERNAL &&
qhwc::needsScaling(layer))
return false;
}
mCurrentFrame.fbCount = 0;
mCurrentFrame.fbZ = -1;
memcpy(&mCurrentFrame.isFBComposed, &mCurrentFrame.drop,
sizeof(mCurrentFrame.isFBComposed));
mCurrentFrame.mdpCount = mCurrentFrame.layerCount - mCurrentFrame.fbCount -
mCurrentFrame.dropCount;
if(sEnable4k2kYUVSplit){
adjustForSourceSplit(ctx, list);
}
if(!resourceCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: resource check failed", __FUNCTION__);
return false;
}
return true;
}
bool MDPComp::partialMDPComp(hwc_context_t *ctx, hwc_display_contents_1_t* list)
{
if(!sEnableMixedMode) {
//Mixed mode is disabled. No need to even try caching.
return false;
}
bool ret = false;
if(isLoadBasedCompDoable(ctx, list)) {
ret = loadBasedCompPreferGPU(ctx, list) ||
loadBasedCompPreferMDP(ctx, list);
}
if(!ret) {
ret = cacheBasedComp(ctx, list);
}
return ret;
}
bool MDPComp::cacheBasedComp(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
mCurrentFrame.reset(numAppLayers);
updateLayerCache(ctx, list);
//If an MDP marked layer is unsupported cannot do partial MDP Comp
for(int i = 0; i < numAppLayers; i++) {
if(!mCurrentFrame.isFBComposed[i]) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if(not isSupportedForMDPComp(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: Unsupported layer in list",
__FUNCTION__);
return false;
}
}
}
updateYUV(ctx, list, false /*secure only*/);
bool ret = markLayersForCaching(ctx, list); //sets up fbZ also
if(!ret) {
ALOGD_IF(isDebug(),"%s: batching failed, dpy %d",__FUNCTION__, mDpy);
return false;
}
int mdpCount = mCurrentFrame.mdpCount;
if(sEnable4k2kYUVSplit){
adjustForSourceSplit(ctx, list);
}
//Will benefit cases where a video has non-updating background.
if((mDpy > HWC_DISPLAY_PRIMARY) and
(mdpCount > MAX_SEC_LAYERS)) {
ALOGD_IF(isDebug(), "%s: Exceeds max secondary pipes",__FUNCTION__);
return false;
}
if(!resourceCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: resource check failed", __FUNCTION__);
return false;
}
return true;
}
bool MDPComp::loadBasedCompPreferGPU(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
mCurrentFrame.reset(numAppLayers);
int stagesForMDP = min(sMaxPipesPerMixer, ctx->mOverlay->availablePipes(
mDpy, Overlay::MIXER_DEFAULT));
//If MDP has X possible stages, it can take X layers.
const int batchSize = numAppLayers - (stagesForMDP - 1); //1 for FB
if(batchSize <= 0) {
ALOGD_IF(isDebug(), "%s: Not attempting", __FUNCTION__);
return false;
}
int minBatchStart = -1;
size_t minBatchPixelCount = SIZE_MAX;
for(int i = 0; i <= numAppLayers - batchSize; i++) {
uint32_t batchPixelCount = 0;
for(int j = i; j < i + batchSize; j++) {
hwc_layer_1_t* layer = &list->hwLayers[j];
hwc_rect_t crop = integerizeSourceCrop(layer->sourceCropf);
batchPixelCount += (crop.right - crop.left) *
(crop.bottom - crop.top);
}
if(batchPixelCount < minBatchPixelCount) {
minBatchPixelCount = batchPixelCount;
minBatchStart = i;
}
}
if(minBatchStart < 0) {
ALOGD_IF(isDebug(), "%s: No batch found batchSize %d numAppLayers %d",
__FUNCTION__, batchSize, numAppLayers);
return false;
}
for(int i = 0; i < numAppLayers; i++) {
if(i < minBatchStart || i >= minBatchStart + batchSize) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if(not isSupportedForMDPComp(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: MDP unsupported layer found at %d",
__FUNCTION__, i);
return false;
}
mCurrentFrame.isFBComposed[i] = false;
}
}
mCurrentFrame.fbZ = minBatchStart;
mCurrentFrame.fbCount = batchSize;
mCurrentFrame.mdpCount = mCurrentFrame.layerCount - batchSize;
if(sEnable4k2kYUVSplit){
adjustForSourceSplit(ctx, list);
}
if(!resourceCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: resource check failed", __FUNCTION__);
return false;
}
ALOGD_IF(isDebug(), "%s: fbZ %d batchSize %d",
__FUNCTION__, mCurrentFrame.fbZ, batchSize);
return true;
}
bool MDPComp::loadBasedCompPreferMDP(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
const int numAppLayers = ctx->listStats[mDpy].numAppLayers;
//TODO get the ib from sysfs node.
//Full screen is from ib perspective, not actual full screen
const int bpp = 4;
double panelRefRate =
1000000000.0 / ctx->dpyAttr[mDpy].vsync_period;
double bwLeft = sMaxBw - sBwClaimed;
const int fullScreenLayers = bwLeft * 1000000000 / (ctx->dpyAttr[mDpy].xres
* ctx->dpyAttr[mDpy].yres * bpp * panelRefRate);
const int fbBatchSize = numAppLayers - (fullScreenLayers - 1);
//If batch size is not at least 2, we aren't really preferring MDP, since
//only 1 layer going to GPU could actually translate into an entire FB
//needed to be fetched by MDP, thus needing more b/w rather than less.
if(fbBatchSize < 2 || fbBatchSize > numAppLayers) {
ALOGD_IF(isDebug(), "%s: Not attempting", __FUNCTION__);
return false;
}
//Top-most layers constitute FB batch
const int fbBatchStart = numAppLayers - fbBatchSize;
//Bottom-most layers constitute MDP batch
for(int i = 0; i < fbBatchStart; i++) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if(not isSupportedForMDPComp(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: MDP unsupported layer found at %d",
__FUNCTION__, i);
return false;
}
mCurrentFrame.isFBComposed[i] = false;
}
mCurrentFrame.fbZ = fbBatchStart;
mCurrentFrame.fbCount = fbBatchSize;
mCurrentFrame.mdpCount = mCurrentFrame.layerCount - fbBatchSize;
if(sEnable4k2kYUVSplit){
adjustForSourceSplit(ctx, list);
}
if(!resourceCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: resource check failed", __FUNCTION__);
return false;
}
ALOGD_IF(isDebug(), "%s: FB Z %d, num app layers %d, MDP Batch Size %d",
__FUNCTION__, mCurrentFrame.fbZ, numAppLayers,
numAppLayers - fbBatchSize);
return true;
}
bool MDPComp::isLoadBasedCompDoable(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
if(mDpy or isSecurePresent(ctx, mDpy) or
not (list->flags & HWC_GEOMETRY_CHANGED)) {
return false;
}
return true;
}
bool MDPComp::isOnlyVideoDoable(hwc_context_t *ctx,
hwc_display_contents_1_t* list, bool secureOnly) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
mCurrentFrame.reset(numAppLayers);
updateYUV(ctx, list, secureOnly);
int mdpCount = mCurrentFrame.mdpCount;
if(!isYuvPresent(ctx, mDpy)) {
return false;
}
/* Bail out if we are processing only secured video layers
* and we dont have any */
if(!isSecurePresent(ctx, mDpy) && secureOnly){
return false;
}
if(!mdpCount)
return false;
if(!resourceCheck(ctx, list)) {
ALOGD_IF(isDebug(), "%s: resource check failed", __FUNCTION__);
return false;
}
return true;
}
/* Checks for conditions where YUV layers cannot be bypassed */
bool MDPComp::isYUVDoable(hwc_context_t* ctx, hwc_layer_1_t* layer) {
bool extAnimBlockFeature = mDpy && ctx->listStats[mDpy].isDisplayAnimating;
if(isSkipLayer(layer) && !extAnimBlockFeature) {
ALOGD_IF(isDebug(), "%s: Video marked SKIP dpy %d", __FUNCTION__, mDpy);
return false;
}
if(layer->transform & HWC_TRANSFORM_ROT_90 && !canUseRotator(ctx,mDpy)) {
ALOGD_IF(isDebug(), "%s: no free DMA pipe",__FUNCTION__);
return false;
}
if(isSecuring(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: MDP securing is active", __FUNCTION__);
return false;
}
if(!isValidDimension(ctx, layer)) {
ALOGD_IF(isDebug(), "%s: Buffer is of invalid width",
__FUNCTION__);
return false;
}
if(layer->planeAlpha < 0xFF) {
ALOGD_IF(isDebug(), "%s: Cannot handle YUV layer with plane alpha\
in video only mode",
__FUNCTION__);
return false;
}
return true;
}
/* starts at fromIndex and check for each layer to find
* if it it has overlapping with any Updating layer above it in zorder
* till the end of the batch. returns true if it finds any intersection */
bool MDPComp::canPushBatchToTop(const hwc_display_contents_1_t* list,
int fromIndex, int toIndex) {
for(int i = fromIndex; i < toIndex; i++) {
if(mCurrentFrame.isFBComposed[i] && !mCurrentFrame.drop[i]) {
if(intersectingUpdatingLayers(list, i+1, toIndex, i)) {
return false;
}
}
}
return true;
}
/* Checks if given layer at targetLayerIndex has any
* intersection with all the updating layers in beween
* fromIndex and toIndex. Returns true if it finds intersectiion */
bool MDPComp::intersectingUpdatingLayers(const hwc_display_contents_1_t* list,
int fromIndex, int toIndex, int targetLayerIndex) {
for(int i = fromIndex; i <= toIndex; i++) {
if(!mCurrentFrame.isFBComposed[i]) {
if(areLayersIntersecting(&list->hwLayers[i],
&list->hwLayers[targetLayerIndex])) {
return true;
}
}
}
return false;
}
int MDPComp::getBatch(hwc_display_contents_1_t* list,
int& maxBatchStart, int& maxBatchEnd,
int& maxBatchCount) {
int i = 0;
int fbZOrder =-1;
while (i < mCurrentFrame.layerCount) {
int batchCount = 0;
int batchStart = i;
int batchEnd = i;
int fbZ = batchStart;
int firstZReverseIndex = -1;
int updatingLayersAbove = 0;//Updating layer count in middle of batch
while(i < mCurrentFrame.layerCount) {
if(!mCurrentFrame.isFBComposed[i]) {
if(!batchCount) {
i++;
break;
}
updatingLayersAbove++;
i++;
continue;
} else {
if(mCurrentFrame.drop[i]) {
i++;
continue;
} else if(updatingLayersAbove <= 0) {
batchCount++;
batchEnd = i;
i++;
continue;
} else { //Layer is FBComposed, not a drop & updatingLayer > 0
// We have a valid updating layer already. If layer-i not
// have overlapping with all updating layers in between
// batch-start and i, then we can add layer i to batch.
if(!intersectingUpdatingLayers(list, batchStart, i-1, i)) {
batchCount++;
batchEnd = i;
i++;
continue;
} else if(canPushBatchToTop(list, batchStart, i)) {
//If All the non-updating layers with in this batch
//does not have intersection with the updating layers
//above in z-order, then we can safely move the batch to
//higher z-order. Increment fbZ as it is moving up.
if( firstZReverseIndex < 0) {
firstZReverseIndex = i;
}
batchCount++;
batchEnd = i;
fbZ += updatingLayersAbove;
i++;
updatingLayersAbove = 0;
continue;
} else {
//both failed.start the loop again from here.
if(firstZReverseIndex >= 0) {
i = firstZReverseIndex;
}
break;
}
}
}
}
if(batchCount > maxBatchCount) {
maxBatchCount = batchCount;
maxBatchStart = batchStart;
maxBatchEnd = batchEnd;
fbZOrder = fbZ;
}
}
return fbZOrder;
}
bool MDPComp::markLayersForCaching(hwc_context_t* ctx,
hwc_display_contents_1_t* list) {
/* Idea is to keep as many non-updating(cached) layers in FB and
* send rest of them through MDP. This is done in 2 steps.
* 1. Find the maximum contiguous batch of non-updating layers.
* 2. See if we can improve this batch size for caching by adding
* opaque layers around the batch, if they don't have
* any overlapping with the updating layers in between.
* NEVER mark an updating layer for caching.
* But cached ones can be marked for MDP */
int maxBatchStart = -1;
int maxBatchEnd = -1;
int maxBatchCount = 0;
int fbZ = -1;
/* All or Nothing is cached. No batching needed */
if(!mCurrentFrame.fbCount) {
mCurrentFrame.fbZ = -1;
return true;
}
if(!mCurrentFrame.mdpCount) {
mCurrentFrame.fbZ = 0;
return true;
}
fbZ = getBatch(list, maxBatchStart, maxBatchEnd, maxBatchCount);
/* reset rest of the layers lying inside ROI for MDP comp */
for(int i = 0; i < mCurrentFrame.layerCount; i++) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if((i < maxBatchStart || i > maxBatchEnd) &&
mCurrentFrame.isFBComposed[i]){
if(!mCurrentFrame.drop[i]){
//If an unsupported layer is being attempted to
//be pulled out we should fail
if(not isSupportedForMDPComp(ctx, layer)) {
return false;
}
mCurrentFrame.isFBComposed[i] = false;
}
}
}
// update the frame data
mCurrentFrame.fbZ = fbZ;
mCurrentFrame.fbCount = maxBatchCount;
mCurrentFrame.mdpCount = mCurrentFrame.layerCount -
mCurrentFrame.fbCount - mCurrentFrame.dropCount;
ALOGD_IF(isDebug(),"%s: cached count: %d",__FUNCTION__,
mCurrentFrame.fbCount);
return true;
}
void MDPComp::updateLayerCache(hwc_context_t* ctx,
hwc_display_contents_1_t* list) {
int numAppLayers = ctx->listStats[mDpy].numAppLayers;
int fbCount = 0;
for(int i = 0; i < numAppLayers; i++) {
hwc_layer_1_t* layer = &list->hwLayers[i];
if (mCachedFrame.hnd[i] == list->hwLayers[i].handle) {
if(!mCurrentFrame.drop[i])
fbCount++;
mCurrentFrame.isFBComposed[i] = true;
} else {
mCurrentFrame.isFBComposed[i] = false;
}
}
mCurrentFrame.fbCount = fbCount;
mCurrentFrame.mdpCount = mCurrentFrame.layerCount - mCurrentFrame.fbCount
- mCurrentFrame.dropCount;
ALOGD_IF(isDebug(),"%s: MDP count: %d FB count %d drop count: %d"
,__FUNCTION__, mCurrentFrame.mdpCount, mCurrentFrame.fbCount,
mCurrentFrame.dropCount);
}
void MDPComp::updateYUV(hwc_context_t* ctx, hwc_display_contents_1_t* list,
bool secureOnly) {
int nYuvCount = ctx->listStats[mDpy].yuvCount;
if(!nYuvCount && mDpy) {
//Reset "No animation on external display" related parameters.
ctx->mPrevCropVideo.left = ctx->mPrevCropVideo.top =
ctx->mPrevCropVideo.right = ctx->mPrevCropVideo.bottom = 0;
ctx->mPrevDestVideo.left = ctx->mPrevDestVideo.top =
ctx->mPrevDestVideo.right = ctx->mPrevDestVideo.bottom = 0;
ctx->mPrevTransformVideo = 0;
return;
}
for(int index = 0;index < nYuvCount; index++){
int nYuvIndex = ctx->listStats[mDpy].yuvIndices[index];
hwc_layer_1_t* layer = &list->hwLayers[nYuvIndex];
if(!isYUVDoable(ctx, layer)) {
if(!mCurrentFrame.isFBComposed[nYuvIndex]) {
mCurrentFrame.isFBComposed[nYuvIndex] = true;
mCurrentFrame.fbCount++;
}
} else {
if(mCurrentFrame.isFBComposed[nYuvIndex]) {
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(!secureOnly || isSecureBuffer(hnd)) {
mCurrentFrame.isFBComposed[nYuvIndex] = false;
mCurrentFrame.fbCount--;
}
}
}
}
mCurrentFrame.mdpCount = mCurrentFrame.layerCount -
mCurrentFrame.fbCount - mCurrentFrame.dropCount;
ALOGD_IF(isDebug(),"%s: fb count: %d",__FUNCTION__,
mCurrentFrame.fbCount);
}
bool MDPComp::programMDP(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
//Configure framebuffer first if applicable
if(mCurrentFrame.fbZ >= 0) {
if(!ctx->mFBUpdate[mDpy]->prepare(ctx, list, mCurrentFrame.fbZ)) {
ALOGD_IF(isDebug(), "%s configure framebuffer failed",
__FUNCTION__);
return false;
}
}
if(!allocLayerPipes(ctx, list)) {
ALOGD_IF(isDebug(), "%s: Unable to allocate MDP pipes", __FUNCTION__);
return false;
}
for (int index = 0, mdpNextZOrder = 0; index < mCurrentFrame.layerCount;
index++) {
if(!mCurrentFrame.isFBComposed[index]) {
int mdpIndex = mCurrentFrame.layerToMDP[index];
hwc_layer_1_t* layer = &list->hwLayers[index];
//Leave fbZ for framebuffer. CACHE/GLES layers go here.
if(mdpNextZOrder == mCurrentFrame.fbZ) {
mdpNextZOrder++;
}
MdpPipeInfo* cur_pipe = mCurrentFrame.mdpToLayer[mdpIndex].pipeInfo;
cur_pipe->zOrder = mdpNextZOrder++;
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit){
if(configure4k2kYuv(ctx, layer,
mCurrentFrame.mdpToLayer[mdpIndex])
!= 0 ){
ALOGD_IF(isDebug(), "%s: Failed to configure split pipes \
for layer %d",__FUNCTION__, index);
return false;
}
else{
mdpNextZOrder++;
}
continue;
}
if(configure(ctx, layer, mCurrentFrame.mdpToLayer[mdpIndex]) != 0 ){
ALOGD_IF(isDebug(), "%s: Failed to configure overlay for \
layer %d",__FUNCTION__, index);
return false;
}
}
}
return true;
}
bool MDPComp::resourceCheck(hwc_context_t *ctx,
hwc_display_contents_1_t *list) {
const bool fbUsed = mCurrentFrame.fbCount;
if(mCurrentFrame.mdpCount > sMaxPipesPerMixer - fbUsed) {
ALOGD_IF(isDebug(), "%s: Exceeds MAX_PIPES_PER_MIXER",__FUNCTION__);
return false;
}
if(!arePipesAvailable(ctx, list)) {
return false;
}
double size = calcMDPBytesRead(ctx, list);
if(!bandwidthCheck(ctx, size)) {
ALOGD_IF(isDebug(), "%s: Exceeds bandwidth",__FUNCTION__);
return false;
}
return true;
}
double MDPComp::calcMDPBytesRead(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
double size = 0;
const double GIG = 1000000000.0;
//Skip for targets where no device tree value for bw is supplied
if(sMaxBw <= 0.0) {
return 0.0;
}
for (uint32_t i = 0; i < list->numHwLayers - 1; i++) {
if(!mCurrentFrame.isFBComposed[i]) {
hwc_layer_1_t* layer = &list->hwLayers[i];
private_handle_t *hnd = (private_handle_t *)layer->handle;
if (hnd) {
hwc_rect_t crop = integerizeSourceCrop(layer->sourceCropf);
hwc_rect_t dst = layer->displayFrame;
float bpp = ((float)hnd->size) / (hnd->width * hnd->height);
size += (bpp * (crop.right - crop.left) *
(crop.bottom - crop.top) *
ctx->dpyAttr[mDpy].yres / (dst.bottom - dst.top)) /
GIG;
}
}
}
if(mCurrentFrame.fbCount) {
hwc_layer_1_t* layer = &list->hwLayers[list->numHwLayers - 1];
int tempw, temph;
size += (getBufferSizeAndDimensions(
layer->displayFrame.right - layer->displayFrame.left,
layer->displayFrame.bottom - layer->displayFrame.top,
HAL_PIXEL_FORMAT_RGBA_8888,
tempw, temph)) / GIG;
}
return size;
}
bool MDPComp::bandwidthCheck(hwc_context_t *ctx, const double& size) {
//Skip for targets where no device tree value for bw is supplied
if(sMaxBw <= 0.0) {
return true;
}
double panelRefRate =
1000000000.0 / ctx->dpyAttr[mDpy].vsync_period;
if((size * panelRefRate) > (sMaxBw - sBwClaimed)) {
return false;
}
return true;
}
bool MDPComp::hwLimitationsCheck(hwc_context_t* ctx,
hwc_display_contents_1_t* list) {
//A-family hw limitation:
//If a layer need alpha scaling, MDP can not support.
if(ctx->mMDP.version < qdutils::MDSS_V5) {
for(int i = 0; i < mCurrentFrame.layerCount; ++i) {
if(!mCurrentFrame.isFBComposed[i] &&
isAlphaScaled( &list->hwLayers[i])) {
ALOGD_IF(isDebug(), "%s:frame needs alphaScaling",__FUNCTION__);
return false;
}
}
}
// On 8x26 & 8974 hw, we have a limitation of downscaling+blending.
//If multiple layers requires downscaling and also they are overlapping
//fall back to GPU since MDSS can not handle it.
if(qdutils::MDPVersion::getInstance().is8x74v2() ||
qdutils::MDPVersion::getInstance().is8x26()) {
for(int i = 0; i < mCurrentFrame.layerCount-1; ++i) {
hwc_layer_1_t* botLayer = &list->hwLayers[i];
if(!mCurrentFrame.isFBComposed[i] &&
isDownscaleRequired(botLayer)) {
//if layer-i is marked for MDP and needs downscaling
//check if any MDP layer on top of i & overlaps with layer-i
for(int j = i+1; j < mCurrentFrame.layerCount; ++j) {
hwc_layer_1_t* topLayer = &list->hwLayers[j];
if(!mCurrentFrame.isFBComposed[j] &&
isDownscaleRequired(topLayer)) {
hwc_rect_t r = getIntersection(botLayer->displayFrame,
topLayer->displayFrame);
if(isValidRect(r))
return false;
}
}
}
}
}
return true;
}
int MDPComp::prepare(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
int ret = 0;
const int numLayers = ctx->listStats[mDpy].numAppLayers;
MDPVersion& mdpVersion = qdutils::MDPVersion::getInstance();
//reset old data
mCurrentFrame.reset(numLayers);
memset(&mCurrentFrame.drop, 0, sizeof(mCurrentFrame.drop));
mCurrentFrame.dropCount = 0;
//number of app layers exceeds MAX_NUM_APP_LAYERS fall back to GPU
//do not cache the information for next draw cycle.
if(numLayers > MAX_NUM_APP_LAYERS) {
mCachedFrame.updateCounts(mCurrentFrame);
ALOGI("%s: Number of App layers exceeded the limit ",
__FUNCTION__);
ret = -1;
return ret;
}
//Hard conditions, if not met, cannot do MDP comp
if(!isFrameDoable(ctx)) {
ALOGD_IF( isDebug(),"%s: MDP Comp not possible for this frame",
__FUNCTION__);
reset(numLayers, list);
ret = -1;
goto exit;
}
generateROI(ctx, list);
//Convert from kbps to gbps
sMaxBw = mdpVersion.getHighBw() / 1000000.0;
if (ctx->mExtDisplay->isConnected() || ctx->mMDP.panel != MIPI_CMD_PANEL) {
sMaxBw = mdpVersion.getLowBw() / 1000000.0;
}
//Check whether layers marked for MDP Composition is actually doable.
if(isFullFrameDoable(ctx, list)) {
mCurrentFrame.map();
//Acquire and Program MDP pipes
if(!programMDP(ctx, list)) {
reset(numLayers, list);
ctx->mOverlay->clear(mDpy);
ret = -1;
goto exit;
} else { //Success
//Any change in composition types needs an FB refresh
mCurrentFrame.needsRedraw = false;
if(!mCachedFrame.isSameFrame(mCurrentFrame, list) ||
(list->flags & HWC_GEOMETRY_CHANGED) ||
isSkipPresent(ctx, mDpy)) {
mCurrentFrame.needsRedraw = true;
}
}
} else if(isOnlyVideoDoable(ctx, list, false /*secure only*/) ||
isOnlyVideoDoable(ctx, list, true /*secure only*/)) {
//All layers marked for MDP comp cannot be bypassed.
//Try to compose atleast YUV layers through MDP comp and let
//all the RGB layers compose in FB
//Destination over
mCurrentFrame.fbZ = -1;
if(mCurrentFrame.fbCount)
mCurrentFrame.fbZ = mCurrentFrame.mdpCount;
if(sEnable4k2kYUVSplit){
adjustForSourceSplit(ctx, list);
}
mCurrentFrame.map();
if(!programMDP(ctx, list)) {
reset(numLayers, list);
ctx->mOverlay->clear(mDpy);
ret = -1;
goto exit;
}
} else {
reset(numLayers, list);
memset(&mCurrentFrame.drop, 0, sizeof(mCurrentFrame.drop));
mCurrentFrame.dropCount = 0;
ret = -1;
goto exit;
}
//UpdateLayerFlags
setMDPCompLayerFlags(ctx, list);
mCachedFrame.cacheAll(list);
mCachedFrame.updateCounts(mCurrentFrame);
// unlock it before calling dump function to avoid deadlock
if(isDebug()) {
ALOGD("GEOMETRY change: %d", (list->flags & HWC_GEOMETRY_CHANGED));
android::String8 sDump("");
dump(sDump);
ALOGE("%s",sDump.string());
}
exit:
double panelRefRate =
1000000000.0 / ctx->dpyAttr[mDpy].vsync_period;
sBwClaimed += calcMDPBytesRead(ctx, list) * panelRefRate;
return ret;
}
bool MDPComp::allocSplitVGPipesfor4k2k(hwc_context_t *ctx,
hwc_display_contents_1_t* list, int index) {
bool bRet = true;
hwc_layer_1_t* layer = &list->hwLayers[index];
private_handle_t *hnd = (private_handle_t *)layer->handle;
int mdpIndex = mCurrentFrame.layerToMDP[index];
PipeLayerPair& info = mCurrentFrame.mdpToLayer[mdpIndex];
info.pipeInfo = new MdpYUVPipeInfo;
info.rot = NULL;
MdpYUVPipeInfo& pipe_info = *(MdpYUVPipeInfo*)info.pipeInfo;
ePipeType type = MDPCOMP_OV_VG;
pipe_info.lIndex = ovutils::OV_INVALID;
pipe_info.rIndex = ovutils::OV_INVALID;
pipe_info.lIndex = getMdpPipe(ctx, type, Overlay::MIXER_DEFAULT);
if(pipe_info.lIndex == ovutils::OV_INVALID){
bRet = false;
ALOGD_IF(isDebug(),"%s: allocating first VG pipe failed",
__FUNCTION__);
}
pipe_info.rIndex = getMdpPipe(ctx, type, Overlay::MIXER_DEFAULT);
if(pipe_info.rIndex == ovutils::OV_INVALID){
bRet = false;
ALOGD_IF(isDebug(),"%s: allocating second VG pipe failed",
__FUNCTION__);
}
return bRet;
}
//=============MDPCompNonSplit===================================================
void MDPCompNonSplit::adjustForSourceSplit(hwc_context_t *ctx,
hwc_display_contents_1_t* list){
//As we split 4kx2k yuv layer and program to 2 VG pipes
//(if available) increase mdpcount accordingly
mCurrentFrame.mdpCount += ctx->listStats[mDpy].yuv4k2kCount;
//If 4k2k Yuv layer split is possible, and if
//fbz is above 4k2k layer, increment fb zorder by 1
//as we split 4k2k layer and increment zorder for right half
//of the layer
if(mCurrentFrame.fbZ >= 0) {
int n4k2kYuvCount = ctx->listStats[mDpy].yuv4k2kCount;
for(int index = 0; index < n4k2kYuvCount; index++){
int n4k2kYuvIndex =
ctx->listStats[mDpy].yuv4k2kIndices[index];
if(mCurrentFrame.fbZ > n4k2kYuvIndex){
mCurrentFrame.fbZ += 1;
}
}
}
}
/*
* Configures pipe(s) for MDP composition
*/
int MDPCompNonSplit::configure(hwc_context_t *ctx, hwc_layer_1_t *layer,
PipeLayerPair& PipeLayerPair) {
MdpPipeInfoNonSplit& mdp_info =
*(static_cast<MdpPipeInfoNonSplit*>(PipeLayerPair.pipeInfo));
eMdpFlags mdpFlags = OV_MDP_BACKEND_COMPOSITION;
eZorder zOrder = static_cast<eZorder>(mdp_info.zOrder);
eIsFg isFg = IS_FG_OFF;
eDest dest = mdp_info.index;
ALOGD_IF(isDebug(),"%s: configuring: layer: %p z_order: %d dest_pipe: %d",
__FUNCTION__, layer, zOrder, dest);
return configureNonSplit(ctx, layer, mDpy, mdpFlags, zOrder, isFg, dest,
&PipeLayerPair.rot);
}
bool MDPCompNonSplit::arePipesAvailable(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
overlay::Overlay& ov = *ctx->mOverlay;
int numPipesNeeded = mCurrentFrame.mdpCount;
int availPipes = ov.availablePipes(mDpy, Overlay::MIXER_DEFAULT);
//Reserve pipe for FB
if(mCurrentFrame.fbCount)
availPipes -= 1;
if(numPipesNeeded > availPipes) {
ALOGD_IF(isDebug(), "%s: Insufficient pipes, dpy %d needed %d, avail %d",
__FUNCTION__, mDpy, numPipesNeeded, availPipes);
return false;
}
if(not areVGPipesAvailable(ctx, list)) {
return false;
}
return true;
}
bool MDPCompNonSplit::areVGPipesAvailable(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
overlay::Overlay& ov = *ctx->mOverlay;
int pipesNeeded = 0;
for(int i = 0; i < mCurrentFrame.layerCount; ++i) {
if(!mCurrentFrame.isFBComposed[i]) {
hwc_layer_1_t* layer = &list->hwLayers[i];
hwc_rect_t dst = layer->displayFrame;
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit){
pipesNeeded = pipesNeeded + 2;
}
else if(isYuvBuffer(hnd)) {
pipesNeeded++;
}
}
}
int availableVGPipes = ov.availablePipes(mDpy, ovutils::OV_MDP_PIPE_VG);
if(pipesNeeded > availableVGPipes) {
ALOGD_IF(isDebug(), "%s: Insufficient VG pipes for video layers"
"dpy %d needed %d, avail %d",
__FUNCTION__, mDpy, pipesNeeded, availableVGPipes);
return false;
}
return true;
}
bool MDPCompNonSplit::allocLayerPipes(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
for(int index = 0; index < mCurrentFrame.layerCount; index++) {
if(mCurrentFrame.isFBComposed[index]) continue;
hwc_layer_1_t* layer = &list->hwLayers[index];
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit){
if(allocSplitVGPipesfor4k2k(ctx, list, index)){
continue;
}
}
int mdpIndex = mCurrentFrame.layerToMDP[index];
PipeLayerPair& info = mCurrentFrame.mdpToLayer[mdpIndex];
info.pipeInfo = new MdpPipeInfoNonSplit;
info.rot = NULL;
MdpPipeInfoNonSplit& pipe_info = *(MdpPipeInfoNonSplit*)info.pipeInfo;
ePipeType type = MDPCOMP_OV_ANY;
if(isYuvBuffer(hnd)) {
type = MDPCOMP_OV_VG;
} else if(!qhwc::needsScaling(layer)
&& Overlay::getDMAMode() != Overlay::DMA_BLOCK_MODE
&& ctx->mMDP.version >= qdutils::MDSS_V5) {
type = MDPCOMP_OV_DMA;
}
pipe_info.index = getMdpPipe(ctx, type, Overlay::MIXER_DEFAULT);
if(pipe_info.index == ovutils::OV_INVALID) {
ALOGD_IF(isDebug(), "%s: Unable to get pipe type = %d",
__FUNCTION__, (int) type);
return false;
}
}
return true;
}
int MDPCompNonSplit::configure4k2kYuv(hwc_context_t *ctx, hwc_layer_1_t *layer,
PipeLayerPair& PipeLayerPair) {
MdpYUVPipeInfo& mdp_info =
*(static_cast<MdpYUVPipeInfo*>(PipeLayerPair.pipeInfo));
eZorder zOrder = static_cast<eZorder>(mdp_info.zOrder);
eIsFg isFg = IS_FG_OFF;
eMdpFlags mdpFlagsL = OV_MDP_BACKEND_COMPOSITION;
eDest lDest = mdp_info.lIndex;
eDest rDest = mdp_info.rIndex;
return configureSourceSplit(ctx, layer, mDpy, mdpFlagsL, zOrder, isFg,
lDest, rDest, &PipeLayerPair.rot);
}
bool MDPCompNonSplit::draw(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
if(!isEnabled()) {
ALOGD_IF(isDebug(),"%s: MDP Comp not configured", __FUNCTION__);
return true;
}
if(!ctx || !list) {
ALOGE("%s: invalid contxt or list",__FUNCTION__);
return false;
}
if(ctx->listStats[mDpy].numAppLayers > MAX_NUM_APP_LAYERS) {
ALOGD_IF(isDebug(),"%s: Exceeding max layer count", __FUNCTION__);
return true;
}
/* reset Invalidator */
if(idleInvalidator && !sIdleFallBack && mCurrentFrame.mdpCount)
idleInvalidator->markForSleep();
overlay::Overlay& ov = *ctx->mOverlay;
LayerProp *layerProp = ctx->layerProp[mDpy];
int numHwLayers = ctx->listStats[mDpy].numAppLayers;
for(int i = 0; i < numHwLayers && mCurrentFrame.mdpCount; i++ )
{
if(mCurrentFrame.isFBComposed[i]) continue;
hwc_layer_1_t *layer = &list->hwLayers[i];
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(!hnd) {
if (!(layer->flags & HWC_COLOR_FILL)) {
ALOGE("%s handle null", __FUNCTION__);
return false;
}
// No PLAY for Color layer
layerProp[i].mFlags &= ~HWC_MDPCOMP;
continue;
}
int mdpIndex = mCurrentFrame.layerToMDP[i];
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit)
{
MdpYUVPipeInfo& pipe_info =
*(MdpYUVPipeInfo*)mCurrentFrame.mdpToLayer[mdpIndex].pipeInfo;
Rotator *rot = mCurrentFrame.mdpToLayer[mdpIndex].rot;
ovutils::eDest indexL = pipe_info.lIndex;
ovutils::eDest indexR = pipe_info.rIndex;
int fd = hnd->fd;
uint32_t offset = hnd->offset;
if(rot) {
rot->queueBuffer(fd, offset);
fd = rot->getDstMemId();
offset = rot->getDstOffset();
}
if(indexL != ovutils::OV_INVALID) {
ovutils::eDest destL = (ovutils::eDest)indexL;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexL );
if (!ov.queueBuffer(fd, offset, destL)) {
ALOGE("%s: queueBuffer failed for display:%d",
__FUNCTION__, mDpy);
return false;
}
}
if(indexR != ovutils::OV_INVALID) {
ovutils::eDest destR = (ovutils::eDest)indexR;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexR );
if (!ov.queueBuffer(fd, offset, destR)) {
ALOGE("%s: queueBuffer failed for display:%d",
__FUNCTION__, mDpy);
return false;
}
}
}
else{
MdpPipeInfoNonSplit& pipe_info =
*(MdpPipeInfoNonSplit*)mCurrentFrame.mdpToLayer[mdpIndex].pipeInfo;
ovutils::eDest dest = pipe_info.index;
if(dest == ovutils::OV_INVALID) {
ALOGE("%s: Invalid pipe index (%d)", __FUNCTION__, dest);
return false;
}
if(!(layerProp[i].mFlags & HWC_MDPCOMP)) {
continue;
}
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer,
hnd, dest );
int fd = hnd->fd;
uint32_t offset = hnd->offset;
Rotator *rot = mCurrentFrame.mdpToLayer[mdpIndex].rot;
if(rot) {
if(!rot->queueBuffer(fd, offset))
return false;
fd = rot->getDstMemId();
offset = rot->getDstOffset();
}
if (!ov.queueBuffer(fd, offset, dest)) {
ALOGE("%s: queueBuffer failed for display:%d ",
__FUNCTION__, mDpy);
return false;
}
}
layerProp[i].mFlags &= ~HWC_MDPCOMP;
}
return true;
}
//=============MDPCompSplit===================================================
void MDPCompSplit::adjustForSourceSplit(hwc_context_t *ctx,
hwc_display_contents_1_t* list){
//if 4kx2k yuv layer is totally present in either in left half
//or right half then try splitting the yuv layer to avoid decimation
int n4k2kYuvCount = ctx->listStats[mDpy].yuv4k2kCount;
const int lSplit = getLeftSplit(ctx, mDpy);
for(int index = 0; index < n4k2kYuvCount; index++){
int n4k2kYuvIndex = ctx->listStats[mDpy].yuv4k2kIndices[index];
hwc_layer_1_t* layer = &list->hwLayers[n4k2kYuvIndex];
hwc_rect_t dst = layer->displayFrame;
if((dst.left > lSplit) || (dst.right < lSplit)) {
mCurrentFrame.mdpCount += 1;
}
if(mCurrentFrame.fbZ > n4k2kYuvIndex){
mCurrentFrame.fbZ += 1;
}
}
}
int MDPCompSplit::pipesNeeded(hwc_context_t *ctx,
hwc_display_contents_1_t* list,
int mixer) {
int pipesNeeded = 0;
const int xres = ctx->dpyAttr[mDpy].xres;
const int lSplit = getLeftSplit(ctx, mDpy);
for(int i = 0; i < mCurrentFrame.layerCount; ++i) {
if(!mCurrentFrame.isFBComposed[i]) {
hwc_layer_1_t* layer = &list->hwLayers[i];
hwc_rect_t dst = layer->displayFrame;
if(mixer == Overlay::MIXER_LEFT && dst.left < lSplit) {
pipesNeeded++;
} else if(mixer == Overlay::MIXER_RIGHT && dst.right > lSplit) {
pipesNeeded++;
}
}
}
return pipesNeeded;
}
bool MDPCompSplit::arePipesAvailable(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
overlay::Overlay& ov = *ctx->mOverlay;
int totalPipesNeeded = 0;
for(int i = 0; i < Overlay::MIXER_MAX; i++) {
int numPipesNeeded = pipesNeeded(ctx, list, i);
int availPipes = ov.availablePipes(mDpy, i);
//Reserve pipe(s)for FB
if(mCurrentFrame.fbCount)
numPipesNeeded += 1;
totalPipesNeeded += numPipesNeeded;
//Per mixer check.
if(numPipesNeeded > availPipes) {
ALOGD_IF(isDebug(), "%s: Insufficient pipes for "
"dpy %d mixer %d needed %d, avail %d",
__FUNCTION__, mDpy, i, numPipesNeeded, availPipes);
return false;
}
}
//Per display check, since unused pipes can get counted twice.
int totalPipesAvailable = ov.availablePipes(mDpy);
if(totalPipesNeeded > totalPipesAvailable) {
ALOGD_IF(isDebug(), "%s: Insufficient pipes for "
"dpy %d needed %d, avail %d",
__FUNCTION__, mDpy, totalPipesNeeded, totalPipesAvailable);
return false;
}
if(not areVGPipesAvailable(ctx, list)) {
return false;
}
return true;
}
bool MDPCompSplit::areVGPipesAvailable(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
overlay::Overlay& ov = *ctx->mOverlay;
int pipesNeeded = 0;
const int lSplit = getLeftSplit(ctx, mDpy);
for(int i = 0; i < mCurrentFrame.layerCount; ++i) {
if(!mCurrentFrame.isFBComposed[i]) {
hwc_layer_1_t* layer = &list->hwLayers[i];
hwc_rect_t dst = layer->displayFrame;
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit){
if((dst.left > lSplit)||(dst.right < lSplit)){
pipesNeeded = pipesNeeded + 2;
continue;
}
}
if(isYuvBuffer(hnd)) {
if(dst.left < lSplit) {
pipesNeeded++;
}
if(dst.right > lSplit) {
pipesNeeded++;
}
}
}
}
int availableVGPipes = ov.availablePipes(mDpy, ovutils::OV_MDP_PIPE_VG);
if(pipesNeeded > availableVGPipes) {
ALOGD_IF(isDebug(), "%s: Insufficient VG pipes for video layers"
"dpy %d needed %d, avail %d",
__FUNCTION__, mDpy, pipesNeeded, availableVGPipes);
return false;
}
return true;
}
bool MDPCompSplit::acquireMDPPipes(hwc_context_t *ctx, hwc_layer_1_t* layer,
MdpPipeInfoSplit& pipe_info,
ePipeType type) {
const int xres = ctx->dpyAttr[mDpy].xres;
const int lSplit = getLeftSplit(ctx, mDpy);
hwc_rect_t dst = layer->displayFrame;
pipe_info.lIndex = ovutils::OV_INVALID;
pipe_info.rIndex = ovutils::OV_INVALID;
if (dst.left < lSplit) {
pipe_info.lIndex = getMdpPipe(ctx, type, Overlay::MIXER_LEFT);
if(pipe_info.lIndex == ovutils::OV_INVALID)
return false;
}
if(dst.right > lSplit) {
pipe_info.rIndex = getMdpPipe(ctx, type, Overlay::MIXER_RIGHT);
if(pipe_info.rIndex == ovutils::OV_INVALID)
return false;
}
return true;
}
bool MDPCompSplit::allocLayerPipes(hwc_context_t *ctx,
hwc_display_contents_1_t* list) {
for(int index = 0 ; index < mCurrentFrame.layerCount; index++) {
if(mCurrentFrame.isFBComposed[index]) continue;
hwc_layer_1_t* layer = &list->hwLayers[index];
private_handle_t *hnd = (private_handle_t *)layer->handle;
hwc_rect_t dst = layer->displayFrame;
const int lSplit = getLeftSplit(ctx, mDpy);
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit){
if((dst.left > lSplit)||(dst.right < lSplit)){
if(allocSplitVGPipesfor4k2k(ctx, list, index)){
continue;
}
}
}
int mdpIndex = mCurrentFrame.layerToMDP[index];
PipeLayerPair& info = mCurrentFrame.mdpToLayer[mdpIndex];
info.pipeInfo = new MdpPipeInfoSplit;
info.rot = NULL;
MdpPipeInfoSplit& pipe_info = *(MdpPipeInfoSplit*)info.pipeInfo;
ePipeType type = MDPCOMP_OV_ANY;
if(isYuvBuffer(hnd)) {
type = MDPCOMP_OV_VG;
} else if(!qhwc::needsScalingWithSplit(ctx, layer, mDpy)
&& Overlay::getDMAMode() != Overlay::DMA_BLOCK_MODE
&& ctx->mMDP.version >= qdutils::MDSS_V5) {
type = MDPCOMP_OV_DMA;
}
if(!acquireMDPPipes(ctx, layer, pipe_info, type)) {
ALOGD_IF(isDebug(), "%s: Unable to get pipe for type = %d",
__FUNCTION__, (int) type);
return false;
}
}
return true;
}
int MDPCompSplit::configure4k2kYuv(hwc_context_t *ctx, hwc_layer_1_t *layer,
PipeLayerPair& PipeLayerPair) {
const int lSplit = getLeftSplit(ctx, mDpy);
hwc_rect_t dst = layer->displayFrame;
if((dst.left > lSplit)||(dst.right < lSplit)){
MdpYUVPipeInfo& mdp_info =
*(static_cast<MdpYUVPipeInfo*>(PipeLayerPair.pipeInfo));
eZorder zOrder = static_cast<eZorder>(mdp_info.zOrder);
eIsFg isFg = IS_FG_OFF;
eMdpFlags mdpFlagsL = OV_MDP_BACKEND_COMPOSITION;
eDest lDest = mdp_info.lIndex;
eDest rDest = mdp_info.rIndex;
return configureSourceSplit(ctx, layer, mDpy, mdpFlagsL, zOrder, isFg,
lDest, rDest, &PipeLayerPair.rot);
}
else{
return configure(ctx, layer, PipeLayerPair);
}
}
/*
* Configures pipe(s) for MDP composition
*/
int MDPCompSplit::configure(hwc_context_t *ctx, hwc_layer_1_t *layer,
PipeLayerPair& PipeLayerPair) {
MdpPipeInfoSplit& mdp_info =
*(static_cast<MdpPipeInfoSplit*>(PipeLayerPair.pipeInfo));
eZorder zOrder = static_cast<eZorder>(mdp_info.zOrder);
eIsFg isFg = IS_FG_OFF;
eMdpFlags mdpFlagsL = OV_MDP_BACKEND_COMPOSITION;
eDest lDest = mdp_info.lIndex;
eDest rDest = mdp_info.rIndex;
ALOGD_IF(isDebug(),"%s: configuring: layer: %p z_order: %d dest_pipeL: %d"
"dest_pipeR: %d",__FUNCTION__, layer, zOrder, lDest, rDest);
return configureSplit(ctx, layer, mDpy, mdpFlagsL, zOrder, isFg, lDest,
rDest, &PipeLayerPair.rot);
}
bool MDPCompSplit::draw(hwc_context_t *ctx, hwc_display_contents_1_t* list) {
if(!isEnabled()) {
ALOGD_IF(isDebug(),"%s: MDP Comp not configured", __FUNCTION__);
return true;
}
if(!ctx || !list) {
ALOGE("%s: invalid contxt or list",__FUNCTION__);
return false;
}
if(ctx->listStats[mDpy].numAppLayers > MAX_NUM_APP_LAYERS) {
ALOGD_IF(isDebug(),"%s: Exceeding max layer count", __FUNCTION__);
return true;
}
/* reset Invalidator */
if(idleInvalidator && !sIdleFallBack && mCurrentFrame.mdpCount)
idleInvalidator->markForSleep();
overlay::Overlay& ov = *ctx->mOverlay;
LayerProp *layerProp = ctx->layerProp[mDpy];
int numHwLayers = ctx->listStats[mDpy].numAppLayers;
for(int i = 0; i < numHwLayers && mCurrentFrame.mdpCount; i++ )
{
if(mCurrentFrame.isFBComposed[i]) continue;
hwc_layer_1_t *layer = &list->hwLayers[i];
private_handle_t *hnd = (private_handle_t *)layer->handle;
if(!hnd) {
ALOGE("%s handle null", __FUNCTION__);
return false;
}
if(!(layerProp[i].mFlags & HWC_MDPCOMP)) {
continue;
}
int mdpIndex = mCurrentFrame.layerToMDP[i];
if(is4kx2kYuvBuffer(hnd) && sEnable4k2kYUVSplit)
{
MdpYUVPipeInfo& pipe_info =
*(MdpYUVPipeInfo*)mCurrentFrame.mdpToLayer[mdpIndex].pipeInfo;
Rotator *rot = mCurrentFrame.mdpToLayer[mdpIndex].rot;
ovutils::eDest indexL = pipe_info.lIndex;
ovutils::eDest indexR = pipe_info.rIndex;
int fd = hnd->fd;
uint32_t offset = hnd->offset;
if(rot) {
rot->queueBuffer(fd, offset);
fd = rot->getDstMemId();
offset = rot->getDstOffset();
}
if(indexL != ovutils::OV_INVALID) {
ovutils::eDest destL = (ovutils::eDest)indexL;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexL );
if (!ov.queueBuffer(fd, offset, destL)) {
ALOGE("%s: queueBuffer failed for display:%d",
__FUNCTION__, mDpy);
return false;
}
}
if(indexR != ovutils::OV_INVALID) {
ovutils::eDest destR = (ovutils::eDest)indexR;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexR );
if (!ov.queueBuffer(fd, offset, destR)) {
ALOGE("%s: queueBuffer failed for display:%d",
__FUNCTION__, mDpy);
return false;
}
}
}
else{
MdpPipeInfoSplit& pipe_info =
*(MdpPipeInfoSplit*)mCurrentFrame.mdpToLayer[mdpIndex].pipeInfo;
Rotator *rot = mCurrentFrame.mdpToLayer[mdpIndex].rot;
ovutils::eDest indexL = pipe_info.lIndex;
ovutils::eDest indexR = pipe_info.rIndex;
int fd = hnd->fd;
int offset = hnd->offset;
if(ctx->mAD->isModeOn()) {
if(ctx->mAD->draw(ctx, fd, offset)) {
fd = ctx->mAD->getDstFd(ctx);
offset = ctx->mAD->getDstOffset(ctx);
}
}
if(rot) {
rot->queueBuffer(fd, offset);
fd = rot->getDstMemId();
offset = rot->getDstOffset();
}
//************* play left mixer **********
if(indexL != ovutils::OV_INVALID) {
ovutils::eDest destL = (ovutils::eDest)indexL;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexL );
if (!ov.queueBuffer(fd, offset, destL)) {
ALOGE("%s: queueBuffer failed for left mixer",
__FUNCTION__);
return false;
}
}
//************* play right mixer **********
if(indexR != ovutils::OV_INVALID) {
ovutils::eDest destR = (ovutils::eDest)indexR;
ALOGD_IF(isDebug(),"%s: MDP Comp: Drawing layer: %p hnd: %p \
using pipe: %d", __FUNCTION__, layer, hnd, indexR );
if (!ov.queueBuffer(fd, offset, destR)) {
ALOGE("%s: queueBuffer failed for right mixer",
__FUNCTION__);
return false;
}
}
}
layerProp[i].mFlags &= ~HWC_MDPCOMP;
}
return true;
}
}; //namespace