Mercurial > sdl-ios-xcode
view src/main/symbian/EKA2/vectorbuffer.h @ 4216:5b99971a27b4 SDL-1.2
Fixed bug #698
Hans de Goede 2009-02-13 01:10:52 PST
Since the new "glitch free" version of pulseaudio (used in Fedora 10 amongst
others), the sound of SDL using apps (like a simple playmus call) has been
crackling.
While looking in to fixing this I noticed that the current pulseaudio code in
SDL uses pa_simple. However pa_simple uses a thread to pump pulseaudio events
and ipc, given that SDL already has its own thread for audio handling this is
clearly suboptimal, leading to unnecessary context switching IPC, etc. Also
pa_simple does not allow one to implement the WaitAudio() callback for SDL
audiodrivers properly.
Given that my work is mostly a rewrite (although some original pieces remain)
I'm attaching the new .c and .h file, as that is easier to review then the huge
diff.
Let me know if you also want the diff.
This new version has the following features:
-no longer use an additional thread next to the SDL sound thread
-do not crackle with glitch free audio
-when used with a newer pulse, which does glitch free audio, the total latency
is
the same as with the alsa driver
-proper WaitAudio() implementation, saving another mixlen worth of latency
-adds a WaitDone() implementation
This patch has been written in consultancy with Lennart Poetering (the
pulseaudio author) and has been reviewed by him for correct use of the pa API.
author | Sam Lantinga <slouken@libsdl.org> |
---|---|
date | Mon, 21 Sep 2009 09:27:08 +0000 |
parents | e85e65aec22f |
children |
line wrap: on
line source
/* vectorbuffer.cpp yet another circle buffer Markus Mertama */ #ifndef __VECTORBUFFER_H__ #define __VECTORBUFFER_H__ #include<e32std.h> #define VLOG(x) #define VECPANIC(x) VectorPanic(x, __LINE__) void VectorPanic(TInt, TInt); //int DEBUG_INT; NONSHARABLE_CLASS(TNodeBuffer) { public: protected: NONSHARABLE_CLASS(TNode) { public: static TNode* Empty(TUint8* iBuffer); static TNode* New(TNode* aPrev, const TDesC8& aData); const TUint8* Ptr() const; TInt Size() const; inline TNode* Succ(); static void SetSucc(TNode*& aNode); void Terminator(TNode* aNode); private: TNode* iSucc; }; }; inline TNodeBuffer::TNode* TNodeBuffer::TNode::Succ() { return iSucc; } template <TInt C> NONSHARABLE_CLASS(TVectorBuffer) : public TNodeBuffer { public: TVectorBuffer(); TInt Append(const TDesC8& aData); // TInt AppendOverwrite(const TDesC8& aData); TPtrC8 Shift(); TPtrC8 operator[](TInt aIndex) const; TInt Size() const; private: TInt GetRoom(TInt aSize) const; TInt Unreserved() const; private: TNode* iTop; TNode* iBottom; TInt iSize; TUint8 iBuffer[C]; }; template <TInt C> TVectorBuffer<C>::TVectorBuffer() : iSize(0) { Mem::FillZ(iBuffer, C); iTop = TNode::Empty(iBuffer); //these points to buffer iBottom = TNode::Empty(iBuffer); } template<TInt C > TInt TVectorBuffer<C>::Unreserved() const { __ASSERT_DEBUG(iBottom < iBottom->Succ(), VECPANIC(KErrCorrupt)); const TInt bytesbetween = reinterpret_cast<const TUint8*>(iBottom->Succ()) - reinterpret_cast<const TUint8*>(iTop); const TInt topsize = sizeof(TNode); if(bytesbetween > 0) //bytesbetween is room between bottom and top { //therefore free room is subracted from free space const TInt room = C - bytesbetween - topsize; return room; } if(bytesbetween == 0) { if(Size() > 0) return 0; else return C - topsize; } const TInt room = -bytesbetween - topsize; //free is space between pointers return room; } template <TInt C> TInt TVectorBuffer<C>::GetRoom(TInt aSize) const { const TInt bytesnew = sizeof(TNode) + aSize; const TInt room = Unreserved() - bytesnew; return room; } template <TInt C> TInt TVectorBuffer<C>::Append(const TDesC8& aData) //ei ole ok! { const TInt len = aData.Length(); if(GetRoom(len) < 0) { return KErrOverflow; } if(iBottom->Succ()->Ptr() - iBuffer > (C - (len + TInt(sizeof(TNode))))) { VLOG("rc"); // RDebug::Print(_L("vector: append")); TNode* p = TNode::Empty(iBuffer); iBottom->Terminator(p); iBottom = p; return Append(aData); // Append(); // iBottom = TNode::New(p, aData); //just append something into end } //DEBUG_INT++; iBottom = TNode::New(iBottom, aData); iSize += len; return KErrNone; } /* template <TInt C> TInt TVectorBuffer<C>::AppendOverwrite(const TDesC8& aData) //ei ole ok! { while(Append(aData) == KErrOverflow) { if(iTop->Succ() == NULL) { return KErrUnderflow; } //Shift(); //data is lost } return KErrNone; } */ template <TInt C> TPtrC8 TVectorBuffer<C>::Shift() { __ASSERT_ALWAYS(iTop->Succ() != NULL, VECPANIC(KErrUnderflow)); //can never pass-by bottom TNode* node = iTop; iTop = iTop->Succ(); if(iTop > node) { // DEBUG_INT--; iSize -= node->Size(); return TPtrC8(node->Ptr(), node->Size()); } else { // RDebug::Print(_L("vector: shift")); return Shift(); //this happens when buffer is terminated, and data lies in next } } template <TInt C> TInt TVectorBuffer<C>::Size() const { return iSize; } template <TInt C> TPtrC8 TVectorBuffer<C>::operator[](TInt aIndex) const { TInt index = 0; TNode* t = iTop->Size() > 0 ? iTop : iTop->Succ(); //eliminate terminator while(index < aIndex) { TNode* nt = t->Succ(); if(nt < t) { nt = nt->Succ(); } t = nt; if(t->Size() > 0) index++; __ASSERT_ALWAYS(t->Succ() != NULL, VECPANIC(KErrUnderflow)); //can never pass-by bottom } return t->Ptr(); } template <class T, TInt C> NONSHARABLE_CLASS(TVector) : public TVectorBuffer<C * sizeof(T)> { public: TVector(); TInt Append(const T& aData); const T& Shift(); TInt Size() const; const T& operator[](TInt aIndex) const; }; template <class T, TInt C> TVector<T, C>::TVector() : TVectorBuffer<C * sizeof(T)>() { } template <class T, TInt C> TInt TVector<T, C>::Append(const T& aData) { const TPckgC<T> data(aData); return TVectorBuffer<C * sizeof(T)>::Append(data); } template <class T, TInt C> const T& TVector<T, C>::Shift() { const TPtrC8 ptr = TVectorBuffer<C * sizeof(T)>::Shift(); return *(reinterpret_cast<const T*>(ptr.Ptr())); } template <class T, TInt C> TInt TVector<T, C>::Size() const { return TVectorBuffer<C * sizeof(T)>::Size() / sizeof(T); } template <class T, TInt C> const T& TVector<T, C>::operator[](TInt aIndex) const { const TPtrC8 ptr = TVectorBuffer<C * sizeof(T)>::operator[](aIndex); return *(reinterpret_cast<const T*>(ptr.Ptr())); } #endif