/* * Copyright (C) 2009-2012 The Android Open Source Project * * 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 "rsContext.h" #include "rsScriptC.h" #include "rsMatrix4x4.h" #include "rsMatrix3x3.h" #include "rsMatrix2x2.h" #include "rsgApiStructs.h" #if !defined(RS_SERVER) && !defined(RS_COMPATIBILITY_LIB) #include "utils/Timers.h" #endif #include <time.h> using namespace android; using namespace android::renderscript; namespace android { namespace renderscript { ////////////////////////////////////////////////////////////////////////////// // Math routines ////////////////////////////////////////////////////////////////////////////// #if 0 static float SC_sinf_fast(float x) { const float A = 1.0f / (2.0f * M_PI); const float B = -16.0f; const float C = 8.0f; // scale angle for easy argument reduction x *= A; if (fabsf(x) >= 0.5f) { // argument reduction x = x - ceilf(x + 0.5f) + 1.0f; } const float y = B * x * fabsf(x) + C * x; return 0.2215f * (y * fabsf(y) - y) + y; } static float SC_cosf_fast(float x) { x += float(M_PI / 2); const float A = 1.0f / (2.0f * M_PI); const float B = -16.0f; const float C = 8.0f; // scale angle for easy argument reduction x *= A; if (fabsf(x) >= 0.5f) { // argument reduction x = x - ceilf(x + 0.5f) + 1.0f; } const float y = B * x * fabsf(x) + C * x; return 0.2215f * (y * fabsf(y) - y) + y; } #endif ////////////////////////////////////////////////////////////////////////////// // Time routines ////////////////////////////////////////////////////////////////////////////// time_t rsrTime(Context *rsc, time_t *timer) { return time(timer); } tm* rsrLocalTime(Context *rsc, tm *local, time_t *timer) { if (!local) { return nullptr; } // The native localtime function is not thread-safe, so we // have to apply locking for proper behavior in RenderScript. pthread_mutex_lock(&rsc->gLibMutex); tm *tmp = localtime(timer); memcpy(local, tmp, sizeof(int)*9); pthread_mutex_unlock(&rsc->gLibMutex); return local; } int64_t rsrUptimeMillis(Context *rsc) { #ifndef RS_SERVER return nanoseconds_to_milliseconds(systemTime(SYSTEM_TIME_MONOTONIC)); #else return 0; #endif } int64_t rsrUptimeNanos(Context *rsc) { #ifndef RS_SERVER return systemTime(SYSTEM_TIME_MONOTONIC); #else return 0; #endif } float rsrGetDt(Context *rsc, const Script *sc) { #ifndef RS_SERVER int64_t l = sc->mEnviroment.mLastDtTime; sc->mEnviroment.mLastDtTime = systemTime(SYSTEM_TIME_MONOTONIC); return ((float)(sc->mEnviroment.mLastDtTime - l)) / 1.0e9; #else return 0.f; #endif } ////////////////////////////////////////////////////////////////////////////// // ////////////////////////////////////////////////////////////////////////////// static void SetObjectRef(const Context *rsc, const ObjectBase *dst, const ObjectBase *src) { //ALOGE("setObjectRef %p,%p %p", rsc, dst, src); if (src) { CHECK_OBJ(src); src->incSysRef(); } if (dst) { CHECK_OBJ(dst); dst->decSysRef(); } } // Legacy, remove when drivers are updated void rsrClearObject(const Context *rsc, void *dst) { ObjectBase **odst = (ObjectBase **)dst; if (ObjectBase::gDebugReferences) { ALOGE("rsrClearObject %p,%p", odst, *odst); } if (odst[0]) { CHECK_OBJ(odst[0]); odst[0]->decSysRef(); } *odst = nullptr; } void rsrClearObject(rs_object_base *dst) { if (ObjectBase::gDebugReferences) { ALOGE("rsrClearObject %p,%p", dst, dst->p); } if (dst->p) { CHECK_OBJ(dst->p); dst->p->decSysRef(); } dst->p = nullptr; } // Legacy, remove when drivers are updated void rsrClearObject(const Context *rsc, rs_object_base *dst) { rsrClearObject(dst); } // Legacy, remove when drivers are updated void rsrSetObject(const Context *rsc, void *dst, ObjectBase *src) { if (src == nullptr) { rsrClearObject(rsc, dst); return; } ObjectBase **odst = (ObjectBase **)dst; if (ObjectBase::gDebugReferences) { ALOGE("rsrSetObject (base) %p,%p %p", dst, *odst, src); } SetObjectRef(rsc, odst[0], src); src->callUpdateCacheObject(rsc, dst); } void rsrSetObject(const Context *rsc, rs_object_base *dst, const ObjectBase *src) { if (src == nullptr) { rsrClearObject(rsc, dst); return; } ObjectBase **odst = (ObjectBase **)dst; if (ObjectBase::gDebugReferences) { ALOGE("rsrSetObject (base) %p,%p %p", dst, *odst, src); } SetObjectRef(rsc, odst[0], src); src->callUpdateCacheObject(rsc, dst); } // Legacy, remove when drivers are updated bool rsrIsObject(const Context *, ObjectBase* src) { ObjectBase **osrc = (ObjectBase **)src; return osrc != nullptr; } bool rsrIsObject(const Context *rsc, rs_object_base o) { return o.p != nullptr; } uint32_t rsrToClient(Context *rsc, int cmdID, const void *data, int len) { //ALOGE("SC_toClient %i %i %i", cmdID, len); return rsc->sendMessageToClient(data, RS_MESSAGE_TO_CLIENT_USER, cmdID, len, false); } uint32_t rsrToClientBlocking(Context *rsc, int cmdID, const void *data, int len) { //ALOGE("SC_toClientBlocking %i %i", cmdID, len); return rsc->sendMessageToClient(data, RS_MESSAGE_TO_CLIENT_USER, cmdID, len, true); } // Keep these two routines (using non-const void pointers) so that we can // still use existing GPU drivers. uint32_t rsrToClient(Context *rsc, int cmdID, void *data, int len) { return rsrToClient(rsc, cmdID, (const void *)data, len); } uint32_t rsrToClientBlocking(Context *rsc, int cmdID, void *data, int len) { return rsrToClientBlocking(rsc, cmdID, (const void *)data, len); } void rsrAllocationIoSend(Context *rsc, Allocation *src) { src->ioSend(rsc); } void rsrAllocationIoReceive(Context *rsc, Allocation *src) { src->ioReceive(rsc); } void rsrForEach(Context *rsc, Script *target, uint32_t slot, uint32_t numInputs, Allocation **in, Allocation *out, const void *usr, uint32_t usrBytes, const RsScriptCall *call) { target->runForEach(rsc, slot, (const Allocation**)in, numInputs, out, usr, usrBytes, call); } void rsrAllocationSyncAll(Context *rsc, Allocation *a, RsAllocationUsageType usage) { a->syncAll(rsc, usage); } void rsrAllocationCopy1DRange(Context *rsc, Allocation *dstAlloc, uint32_t dstOff, uint32_t dstMip, uint32_t count, Allocation *srcAlloc, uint32_t srcOff, uint32_t srcMip) { rsi_AllocationCopy2DRange(rsc, dstAlloc, dstOff, 0, dstMip, 0, count, 1, srcAlloc, srcOff, 0, srcMip, 0); } void rsrAllocationCopy2DRange(Context *rsc, Allocation *dstAlloc, uint32_t dstXoff, uint32_t dstYoff, uint32_t dstMip, uint32_t dstFace, uint32_t width, uint32_t height, Allocation *srcAlloc, uint32_t srcXoff, uint32_t srcYoff, uint32_t srcMip, uint32_t srcFace) { rsi_AllocationCopy2DRange(rsc, dstAlloc, dstXoff, dstYoff, dstMip, dstFace, width, height, srcAlloc, srcXoff, srcYoff, srcMip, srcFace); } RsElement rsrElementCreate(Context *rsc, RsDataType dt, RsDataKind dk, bool norm, uint32_t vecSize) { return rsi_ElementCreate(rsc, dt, dk, norm, vecSize); } RsType rsrTypeCreate(Context *rsc, const RsElement element, uint32_t dimX, uint32_t dimY, uint32_t dimZ, bool mipmaps, bool faces, uint32_t yuv) { return rsi_TypeCreate(rsc, element, dimX, dimY, dimZ, mipmaps, faces, yuv); } RsAllocation rsrAllocationCreateTyped(Context *rsc, const RsType type, RsAllocationMipmapControl mipmaps, uint32_t usages, uintptr_t ptr) { return rsi_AllocationCreateTyped(rsc, type, mipmaps, usages, ptr); } } }