mirror of
https://github.com/hrydgard/ppsspp.git
synced 2025-04-02 11:01:50 -04:00
Don't queue Flush() etc. while on GPU thread.
Oops, that makes it not actually flush properly of course. Also made a simpler guard wrapper to avoid extra braces in too many places. Reduced contention areas slightly.
This commit is contained in:
Unknown W. Brackets
parent
68fb78e203
commit
e71cf2ec7e
5 changed files with 106 additions and 81 deletions
|
|
@ -74,6 +74,14 @@ volatile CoreState coreState = CORE_STEPPING;
|
|||
volatile bool coreStatePending = false;
|
||||
static volatile CPUThreadState cpuThreadState = CPU_THREAD_NOT_RUNNING;
|
||||
|
||||
bool IsOnSeparateCPUThread() {
|
||||
if (g_Config.bSeparateCPUThread) {
|
||||
return cpuThread->get_id() == std::this_thread::get_id();
|
||||
} else {
|
||||
return false;
|
||||
}
|
||||
}
|
||||
|
||||
bool CPU_NextState(CPUThreadState from, CPUThreadState to) {
|
||||
if (cpuThreadState == from) {
|
||||
cpuThreadState = to;
|
||||
|
|
|
|||
|
|
@ -44,6 +44,8 @@ void PSP_Shutdown();
|
|||
void PSP_RunLoopUntil(u64 globalticks);
|
||||
void PSP_RunLoopFor(int cycles);
|
||||
|
||||
bool IsOnSeparateCPUThread();
|
||||
|
||||
void GetSysDirectories(std::string &memstickpath, std::string &flash0path);
|
||||
|
||||
// RUNNING must be at 0, NEXTFRAME must be at 1.
|
||||
|
|
|
|||
|
|
@ -1121,7 +1121,11 @@ void GLES_GPU::ClearCacheNextFrame() {
|
|||
|
||||
|
||||
void GLES_GPU::Flush() {
|
||||
ScheduleEvent(GPU_EVENT_FLUSH);
|
||||
if (IsOnSeparateCPUThread()) {
|
||||
ScheduleEvent(GPU_EVENT_FLUSH);
|
||||
} else {
|
||||
FlushInternal();
|
||||
}
|
||||
}
|
||||
|
||||
void GLES_GPU::FlushInternal() {
|
||||
|
|
|
|||
|
|
@ -32,7 +32,7 @@ GPUCommon::GPUCommon() :
|
|||
}
|
||||
|
||||
void GPUCommon::PopDLQueue() {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
if(!dlQueue.empty()) {
|
||||
dlQueue.pop_front();
|
||||
if(!dlQueue.empty()) {
|
||||
|
|
@ -47,7 +47,7 @@ void GPUCommon::PopDLQueue() {
|
|||
}
|
||||
|
||||
u32 GPUCommon::DrawSync(int mode) {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
if (mode < 0 || mode > 1)
|
||||
return SCE_KERNEL_ERROR_INVALID_MODE;
|
||||
|
||||
|
|
@ -82,18 +82,16 @@ u32 GPUCommon::DrawSync(int mode) {
|
|||
return PSP_GE_LIST_DRAWING;
|
||||
}
|
||||
|
||||
void GPUCommon::CheckDrawSync()
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
void GPUCommon::CheckDrawSync() {
|
||||
easy_guard guard(listLock);
|
||||
if (dlQueue.empty()) {
|
||||
for (int i = 0; i < DisplayListMaxCount; ++i)
|
||||
dls[i].state = PSP_GE_DL_STATE_NONE;
|
||||
}
|
||||
}
|
||||
|
||||
int GPUCommon::ListSync(int listid, int mode)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
int GPUCommon::ListSync(int listid, int mode) {
|
||||
easy_guard guard(listLock);
|
||||
if (listid < 0 || listid >= DisplayListMaxCount)
|
||||
return SCE_KERNEL_ERROR_INVALID_ID;
|
||||
|
||||
|
|
@ -125,14 +123,14 @@ int GPUCommon::ListSync(int listid, int mode)
|
|||
}
|
||||
|
||||
if (dl.waitTicks > CoreTiming::GetTicks()) {
|
||||
guard.unlock();
|
||||
__KernelWaitCurThread(WAITTYPE_GELISTSYNC, listid, 0, 0, false, "GeListSync");
|
||||
}
|
||||
return PSP_GE_LIST_COMPLETED;
|
||||
}
|
||||
|
||||
u32 GPUCommon::EnqueueList(u32 listpc, u32 stall, int subIntrBase, bool head)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
u32 GPUCommon::EnqueueList(u32 listpc, u32 stall, int subIntrBase, bool head) {
|
||||
easy_guard guard(listLock);
|
||||
// TODO Check the stack values in missing arg and ajust the stack depth
|
||||
|
||||
// Check alignment
|
||||
|
|
@ -216,15 +214,15 @@ u32 GPUCommon::EnqueueList(u32 listpc, u32 stall, int subIntrBase, bool head)
|
|||
drawCompleteTicks = (u64)-1;
|
||||
|
||||
// TODO save context when starting the list if param is set
|
||||
guard.unlock();
|
||||
ProcessDLQueue();
|
||||
}
|
||||
|
||||
return id;
|
||||
}
|
||||
|
||||
u32 GPUCommon::DequeueList(int listid)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
u32 GPUCommon::DequeueList(int listid) {
|
||||
easy_guard guard(listLock);
|
||||
if (listid < 0 || listid >= DisplayListMaxCount || dls[listid].state == PSP_GE_DL_STATE_NONE)
|
||||
return SCE_KERNEL_ERROR_INVALID_ID;
|
||||
|
||||
|
|
@ -239,6 +237,7 @@ u32 GPUCommon::DequeueList(int listid)
|
|||
dlQueue.remove(listid);
|
||||
|
||||
dls[listid].waitTicks = 0;
|
||||
guard.unlock();
|
||||
__KernelTriggerWait(WAITTYPE_GELISTSYNC, listid, 0, "GeListSync");
|
||||
|
||||
CheckDrawSync();
|
||||
|
|
@ -246,9 +245,8 @@ u32 GPUCommon::DequeueList(int listid)
|
|||
return 0;
|
||||
}
|
||||
|
||||
u32 GPUCommon::UpdateStall(int listid, u32 newstall)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
u32 GPUCommon::UpdateStall(int listid, u32 newstall) {
|
||||
easy_guard guard(listLock);
|
||||
if (listid < 0 || listid >= DisplayListMaxCount || dls[listid].state == PSP_GE_DL_STATE_NONE)
|
||||
return SCE_KERNEL_ERROR_INVALID_ID;
|
||||
|
||||
|
|
@ -257,14 +255,14 @@ u32 GPUCommon::UpdateStall(int listid, u32 newstall)
|
|||
if (dls[listid].signal == PSP_GE_SIGNAL_HANDLER_PAUSE)
|
||||
dls[listid].signal = PSP_GE_SIGNAL_HANDLER_SUSPEND;
|
||||
|
||||
guard.unlock();
|
||||
ProcessDLQueue();
|
||||
|
||||
return 0;
|
||||
}
|
||||
|
||||
u32 GPUCommon::Continue()
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
u32 GPUCommon::Continue() {
|
||||
easy_guard guard(listLock);
|
||||
if (!currentList)
|
||||
return 0;
|
||||
|
||||
|
|
@ -300,13 +298,13 @@ u32 GPUCommon::Continue()
|
|||
return -1;
|
||||
}
|
||||
|
||||
guard.unlock();
|
||||
ProcessDLQueue();
|
||||
return 0;
|
||||
}
|
||||
|
||||
u32 GPUCommon::Break(int mode)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
u32 GPUCommon::Break(int mode) {
|
||||
easy_guard guard(listLock);
|
||||
if (mode < 0 || mode > 1)
|
||||
return SCE_KERNEL_ERROR_INVALID_MODE;
|
||||
|
||||
|
|
@ -377,49 +375,47 @@ bool GPUCommon::InterpretList(DisplayList &list) {
|
|||
start = time_now_d();
|
||||
}
|
||||
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
|
||||
// TODO: This has to be right... but it freezes right now?
|
||||
//if (list.state == PSP_GE_DL_STATE_PAUSED)
|
||||
// return false;
|
||||
currentList = &list;
|
||||
// TODO: This has to be right... but it freezes right now?
|
||||
//if (list.state == PSP_GE_DL_STATE_PAUSED)
|
||||
// return false;
|
||||
currentList = &list;
|
||||
|
||||
// I don't know if this is the correct place to zero this, but something
|
||||
// need to do it. See Sol Trigger title screen.
|
||||
// TODO: Maybe this is per list? Should a stalled list remember the old value?
|
||||
gstate_c.offsetAddr = 0;
|
||||
// I don't know if this is the correct place to zero this, but something
|
||||
// need to do it. See Sol Trigger title screen.
|
||||
// TODO: Maybe this is per list? Should a stalled list remember the old value?
|
||||
gstate_c.offsetAddr = 0;
|
||||
|
||||
if (!Memory::IsValidAddress(list.pc)) {
|
||||
ERROR_LOG_REPORT(G3D, "DL PC = %08x WTF!!!!", list.pc);
|
||||
return true;
|
||||
}
|
||||
if (!Memory::IsValidAddress(list.pc)) {
|
||||
ERROR_LOG_REPORT(G3D, "DL PC = %08x WTF!!!!", list.pc);
|
||||
return true;
|
||||
}
|
||||
|
||||
#if defined(USING_QT_UI)
|
||||
if (host->GpuStep()) {
|
||||
host->SendGPUStart();
|
||||
}
|
||||
if (host->GpuStep()) {
|
||||
host->SendGPUStart();
|
||||
}
|
||||
#endif
|
||||
|
||||
cycleLastPC = list.pc;
|
||||
downcount = list.stall == 0 ? 0xFFFFFFF : (list.stall - list.pc) / 4;
|
||||
list.state = PSP_GE_DL_STATE_RUNNING;
|
||||
list.interrupted = false;
|
||||
cycleLastPC = list.pc;
|
||||
downcount = list.stall == 0 ? 0xFFFFFFF : (list.stall - list.pc) / 4;
|
||||
list.state = PSP_GE_DL_STATE_RUNNING;
|
||||
list.interrupted = false;
|
||||
|
||||
gpuState = list.pc == list.stall ? GPUSTATE_STALL : GPUSTATE_RUNNING;
|
||||
}
|
||||
gpuState = list.pc == list.stall ? GPUSTATE_STALL : GPUSTATE_RUNNING;
|
||||
guard.unlock();
|
||||
|
||||
const bool dumpThisFrame = dumpThisFrame_;
|
||||
// TODO: Add check for displaylist debugger.
|
||||
const bool useFastRunLoop = !dumpThisFrame;
|
||||
while (gpuState == GPUSTATE_RUNNING) {
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
if (list.pc == list.stall) {
|
||||
gpuState = GPUSTATE_STALL;
|
||||
downcount = 0;
|
||||
}
|
||||
guard.lock();
|
||||
if (list.pc == list.stall) {
|
||||
gpuState = GPUSTATE_STALL;
|
||||
downcount = 0;
|
||||
}
|
||||
guard.unlock();
|
||||
|
||||
if (useFastRunLoop) {
|
||||
FastRunLoop(list);
|
||||
|
|
@ -427,15 +423,14 @@ bool GPUCommon::InterpretList(DisplayList &list) {
|
|||
SlowRunLoop(list);
|
||||
}
|
||||
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
downcount = list.stall == 0 ? 0xFFFFFFF : (list.stall - list.pc) / 4;
|
||||
guard.lock();
|
||||
downcount = list.stall == 0 ? 0xFFFFFFF : (list.stall - list.pc) / 4;
|
||||
|
||||
if (gpuState == GPUSTATE_STALL && list.stall != list.pc) {
|
||||
// Unstalled.
|
||||
gpuState = GPUSTATE_RUNNING;
|
||||
}
|
||||
if (gpuState == GPUSTATE_STALL && list.stall != list.pc) {
|
||||
// Unstalled.
|
||||
gpuState = GPUSTATE_RUNNING;
|
||||
}
|
||||
guard.unlock();
|
||||
}
|
||||
|
||||
// We haven't run the op at list.pc, so it shouldn't count.
|
||||
|
|
@ -492,7 +487,11 @@ inline void GPUCommon::UpdatePC(u32 currentPC, u32 newPC) {
|
|||
}
|
||||
|
||||
void GPUCommon::ReapplyGfxState() {
|
||||
ScheduleEvent(GPU_EVENT_REAPPLY_GFX_STATE);
|
||||
if (IsOnSeparateCPUThread()) {
|
||||
ScheduleEvent(GPU_EVENT_REAPPLY_GFX_STATE);
|
||||
} else {
|
||||
ReapplyGfxStateInternal();
|
||||
}
|
||||
}
|
||||
|
||||
void GPUCommon::ReapplyGfxStateInternal() {
|
||||
|
|
@ -539,7 +538,7 @@ inline void GPUCommon::UpdateState(GPUState state) {
|
|||
}
|
||||
|
||||
GPUEvent GPUCommon::GetNextEvent() {
|
||||
lock_guard guard(eventsLock);
|
||||
easy_guard guard(eventsLock);
|
||||
if (events.empty()) {
|
||||
eventsDrain.notify_one();
|
||||
return GPU_EVENT_INVALID;
|
||||
|
|
@ -551,14 +550,15 @@ GPUEvent GPUCommon::GetNextEvent() {
|
|||
}
|
||||
|
||||
bool GPUCommon::HasEvents() {
|
||||
lock_guard guard(eventsLock);
|
||||
easy_guard guard(eventsLock);
|
||||
return !events.empty();
|
||||
}
|
||||
|
||||
void GPUCommon::ScheduleEvent(GPUEvent ev) {
|
||||
lock_guard guard(eventsLock);
|
||||
easy_guard guard(eventsLock);
|
||||
events.push_back(ev);
|
||||
eventsWait.notify_one();
|
||||
guard.unlock();
|
||||
|
||||
if (!g_Config.bSeparateCPUThread) {
|
||||
RunEventsUntil(0);
|
||||
|
|
@ -606,7 +606,7 @@ void GPUCommon::SyncThread() {
|
|||
}
|
||||
|
||||
int GPUCommon::GetNextListIndex() {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
auto iter = dlQueue.begin();
|
||||
if (iter != dlQueue.end()) {
|
||||
return *iter;
|
||||
|
|
@ -635,14 +635,15 @@ void GPUCommon::ProcessDLQueueInternal() {
|
|||
if (!InterpretList(l)) {
|
||||
return;
|
||||
} else {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
// At the end, we can remove it from the queue and continue.
|
||||
dlQueue.erase(std::remove(dlQueue.begin(), dlQueue.end(), listIndex), dlQueue.end());
|
||||
}
|
||||
}
|
||||
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
currentList = NULL;
|
||||
guard.unlock();
|
||||
|
||||
drawCompleteTicks = startingTicks + cyclesExecuted;
|
||||
busyTicks = std::max(busyTicks, drawCompleteTicks);
|
||||
|
|
@ -668,14 +669,14 @@ void GPUCommon::ExecuteOp(u32 op, u32 diff) {
|
|||
|
||||
case GE_CMD_ORIGIN:
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
gstate_c.offsetAddr = currentList->pc;
|
||||
}
|
||||
break;
|
||||
|
||||
case GE_CMD_JUMP:
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
u32 target = gstate_c.getRelativeAddress(data);
|
||||
if (Memory::IsValidAddress(target)) {
|
||||
UpdatePC(currentList->pc, target - 4);
|
||||
|
|
@ -688,7 +689,7 @@ void GPUCommon::ExecuteOp(u32 op, u32 diff) {
|
|||
|
||||
case GE_CMD_CALL:
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
// Saint Seiya needs correct support for relative calls.
|
||||
u32 retval = currentList->pc + 4;
|
||||
u32 target = gstate_c.getRelativeAddress(data);
|
||||
|
|
@ -708,7 +709,7 @@ void GPUCommon::ExecuteOp(u32 op, u32 diff) {
|
|||
|
||||
case GE_CMD_RET:
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
if (currentList->stackptr == 0) {
|
||||
ERROR_LOG_REPORT(G3D, "RET: Stack empty!");
|
||||
} else {
|
||||
|
|
@ -731,7 +732,7 @@ void GPUCommon::ExecuteOp(u32 op, u32 diff) {
|
|||
break;
|
||||
|
||||
case GE_CMD_END: {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
u32 prev = Memory::ReadUnchecked_U32(currentList->pc - 4);
|
||||
UpdatePC(currentList->pc);
|
||||
switch (prev >> 24) {
|
||||
|
|
@ -867,7 +868,7 @@ void GPUCommon::ExecuteOp(u32 op, u32 diff) {
|
|||
}
|
||||
|
||||
void GPUCommon::DoState(PointerWrap &p) {
|
||||
lock_guard guard(listLock);
|
||||
easy_guard guard(listLock);
|
||||
|
||||
p.Do<int>(dlQueue);
|
||||
p.DoArray(dls, ARRAY_SIZE(dls));
|
||||
|
|
@ -892,13 +893,11 @@ void GPUCommon::DoState(PointerWrap &p) {
|
|||
p.DoMarker("GPUCommon");
|
||||
}
|
||||
|
||||
void GPUCommon::InterruptStart(int listid)
|
||||
{
|
||||
void GPUCommon::InterruptStart(int listid) {
|
||||
interruptRunning = true;
|
||||
}
|
||||
void GPUCommon::InterruptEnd(int listid)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
void GPUCommon::InterruptEnd(int listid) {
|
||||
easy_guard guard(listLock);
|
||||
interruptRunning = false;
|
||||
isbreak = false;
|
||||
|
||||
|
|
@ -912,13 +911,13 @@ void GPUCommon::InterruptEnd(int listid)
|
|||
if (dl.signal == PSP_GE_SIGNAL_HANDLER_PAUSE)
|
||||
dl.signal = PSP_GE_SIGNAL_HANDLER_SUSPEND;
|
||||
|
||||
guard.unlock();
|
||||
ProcessDLQueue();
|
||||
}
|
||||
|
||||
// TODO: Maybe cleaner to keep this in GE and trigger the clear directly?
|
||||
void GPUCommon::SyncEnd(WaitType waitType, int listid, bool wokeThreads)
|
||||
{
|
||||
lock_guard guard(listLock);
|
||||
void GPUCommon::SyncEnd(WaitType waitType, int listid, bool wokeThreads) {
|
||||
easy_guard guard(listLock);
|
||||
if (waitType == WAITTYPE_GEDRAWSYNC && wokeThreads)
|
||||
{
|
||||
for (int i = 0; i < DisplayListMaxCount; ++i) {
|
||||
|
|
|
|||
|
|
@ -51,6 +51,18 @@ protected:
|
|||
void ReapplyGfxStateInternal();
|
||||
virtual void ProcessEvent(GPUEvent ev) = 0;
|
||||
|
||||
class easy_guard {
|
||||
public:
|
||||
easy_guard(recursive_mutex &mtx) : mtx_(mtx), locked_(true) { mtx_.lock(); }
|
||||
~easy_guard() { if (locked_) mtx_.unlock(); }
|
||||
void lock() { if (!locked_) mtx_.lock(); locked_ = true; }
|
||||
void unlock() { if (locked_) mtx_.unlock(); locked_ = false; }
|
||||
|
||||
private:
|
||||
bool locked_;
|
||||
recursive_mutex &mtx_;
|
||||
};
|
||||
|
||||
typedef std::list<int> DisplayListQueue;
|
||||
|
||||
DisplayList dls[DisplayListMaxCount];
|
||||
|
|
|
|||
Loading…
Add table
Reference in a new issue