Issue #1258 - Part 1: Import mailnews, ldap, and mork from comm-esr52.9.1

1 files changed, 321 insertions, 0 deletions

diff --git a/db/mork/src/morkZone.h b/db/mork/src/morkZone.h
new file mode 100644
index 000000000..9e5ffc54c
--- /dev/null
+++ b/db/mork/src/morkZone.h
@@ -0,0 +1,321 @@
+/* -*- Mode: C++; tab-width: 4; indent-tabs-mode: nil; c-basic-offset: 4 -*-  */
+/* This Source Code Form is subject to the terms of the Mozilla Public
+ * License, v. 2.0. If a copy of the MPL was not distributed with this
+ * file, You can obtain one at http://mozilla.org/MPL/2.0/. */
+
+#ifndef _MORKZONE_
+#define _MORKZONE_ 1
+
+#ifndef _MORK_
+#include "mork.h"
+#endif
+
+#ifndef _MORKNODE_
+#include "morkNode.h"
+#endif
+
+#ifndef _MORKDEQUE_
+#include "morkDeque.h"
+#endif
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+/*| CONFIG_DEBUG: do paranoid debug checks if defined.
+|*/
+#ifdef MORK_DEBUG
+#define morkZone_CONFIG_DEBUG 1 /* debug paranoid if defined */
+#endif /*MORK_DEBUG*/
+
+/*| CONFIG_STATS: keep volume and usage statistics.
+|*/
+#define morkZone_CONFIG_VOL_STATS 1 /* count space used by zone instance */
+
+/*| CONFIG_ARENA: if this is defined, then the morkZone class will alloc big
+**| blocks from the zone's heap, and suballocate from these.  If undefined,
+**| then morkZone will just pass all calls through to the zone's heap.
+|*/
+#ifdef MORK_ENABLE_ZONE_ARENAS
+#define morkZone_CONFIG_ARENA 1 /* be arena, if defined; otherwise no-op */
+#endif /*MORK_ENABLE_ZONE_ARENAS*/
+
+/*| CONFIG_ALIGN_8: if this is defined, then the morkZone class will give
+**| blocks 8 byte alignment instead of only 4 byte alignment.
+|*/
+#ifdef MORK_CONFIG_ALIGN_8
+#define morkZone_CONFIG_ALIGN_8 1 /* ifdef: align to 8 bytes, otherwise 4 */
+#endif /*MORK_CONFIG_ALIGN_8*/
+
+/*| CONFIG_PTR_SIZE_4: if this is defined, then the morkZone class will
+**| assume sizeof(void*) == 4, so a tag slot for padding is needed.
+|*/
+#ifdef MORK_CONFIG_PTR_SIZE_4
+#define morkZone_CONFIG_PTR_SIZE_4 1 /* ifdef: sizeof(void*) == 4 */
+#endif /*MORK_CONFIG_PTR_SIZE_4*/
+
+/*| morkZone_USE_TAG_SLOT: if this is defined, then define slot mRun_Tag
+**| in order to achieve eight byte alignment after the mRun_Next slot.
+|*/
+#if defined(morkZone_CONFIG_ALIGN_8) && defined(morkZone_CONFIG_PTR_SIZE_4)
+#define morkRun_USE_TAG_SLOT 1  /* need mRun_Tag slot inside morkRun */
+#define morkHunk_USE_TAG_SLOT 1 /* need mHunk_Tag slot inside morkHunk */
+#endif
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+#define morkRun_kTag ((mork_u4) 0x6D52754E ) /* ascii 'mRuN' */
+
+/*| morkRun: structure used by morkZone for sized blocks
+|*/
+class morkRun {
+
+protected: // member variable slots
+#ifdef morkRun_USE_TAG_SLOT
+  mork_u4   mRun_Tag; // force 8 byte alignment after mRun_Next
+#endif /* morkRun_USE_TAG_SLOT */
+
+  morkRun*  mRun_Next;
+  
+public: // pointer interpretation of mRun_Next (when inside a list):
+  morkRun*   RunNext() const { return mRun_Next; }
+  void       RunSetNext(morkRun* ioNext) { mRun_Next = ioNext; }
+  
+public: // size interpretation of mRun_Next (when not inside a list):
+  mork_size  RunSize() const { return (mork_size) ((mork_ip) mRun_Next); }
+  void       RunSetSize(mork_size inSize)
+             { mRun_Next = (morkRun*) ((mork_ip) inSize); }
+  
+public: // maintenance and testing of optional tag magic signature slot:
+#ifdef morkRun_USE_TAG_SLOT
+  void       RunInitTag() { mRun_Tag = morkRun_kTag; }
+  mork_bool  RunGoodTag() { return ( mRun_Tag == morkRun_kTag ); }
+#endif /* morkRun_USE_TAG_SLOT */
+  
+public: // conversion back and forth to inline block following run instance:
+  void* RunAsBlock() { return (((mork_u1*) this) + sizeof(morkRun)); }
+  
+  static morkRun* BlockAsRun(void* ioBlock)
+  { return (morkRun*) (((mork_u1*) ioBlock) - sizeof(morkRun)); }
+
+public: // typing & errors
+  static void BadRunTagError(morkEnv* ev);
+  static void RunSizeAlignError(morkEnv* ev);
+};
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+
+/*| morkOldRun: more space to record size when run is put into old free list
+|*/
+class morkOldRun : public morkRun {
+
+protected: // need another size field when mRun_Next is used for linkage:
+  mdb_size mOldRun_Size;
+  
+public: // size getter/setter
+  mork_size  OldSize() const { return mOldRun_Size; }
+  void       OldSetSize(mork_size inSize) { mOldRun_Size = inSize; }
+  
+};
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+#define morkHunk_kTag ((mork_u4) 0x68556E4B ) /* ascii 'hUnK' */
+
+/*| morkHunk: structure used by morkZone for heap allocations.
+|*/
+class morkHunk {
+
+protected: // member variable slots
+
+#ifdef morkHunk_USE_TAG_SLOT
+  mork_u4   mHunk_Tag; // force 8 byte alignment after mHunk_Next
+#endif /* morkHunk_USE_TAG_SLOT */
+
+  morkHunk* mHunk_Next;
+  
+  morkRun   mHunk_Run;
+  
+public: // setters
+  void      HunkSetNext(morkHunk* ioNext) { mHunk_Next = ioNext; }
+  
+public: // getters
+  morkHunk* HunkNext() const { return mHunk_Next; }
+
+  morkRun*  HunkRun() { return &mHunk_Run; }
+  
+public: // maintenance and testing of optional tag magic signature slot:
+#ifdef morkHunk_USE_TAG_SLOT
+  void       HunkInitTag() { mHunk_Tag = morkHunk_kTag; }
+  mork_bool  HunkGoodTag() { return ( mHunk_Tag == morkHunk_kTag ); }
+#endif /* morkHunk_USE_TAG_SLOT */
+
+public: // typing & errors
+  static void BadHunkTagWarning(morkEnv* ev);
+  
+};
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+/*| kNewHunkSize: the default size for a hunk, assuming we must allocate
+**| a new one whenever the free hunk list does not already have.  Note this
+**| number should not be changed without also considering suitable changes
+**| in the related kMaxHunkWaste and kMinHunkSize constants. 
+|*/
+#define morkZone_kNewHunkSize ((mork_size) (64 * 1024)) /* 64K per hunk */
+
+/*| kMaxFreeVolume: some number of bytes of free space in the free hunk list
+**| over which we no longer want to add more free hunks to the list, for fear
+**| of accumulating too much unused, fragmented free space.  This should be a
+**| small multiple of kNewHunkSize, say about two to four times as great, to
+**| allow for no more free hunk space than fits in a handful of new hunks.
+**| This strategy will let us usefuly accumulate "some" free space in the
+**| free hunk list, but without accumulating "too much" free space that way.
+|*/
+#define morkZone_kMaxFreeVolume (morkZone_kNewHunkSize * 3)
+
+/*| kMaxHunkWaste: if a current request is larger than this, and we cannot
+**| satisfy the request with the current hunk, then we just allocate the
+**| block from the heap without changing the current hunk.  Basically this
+**| number represents the largest amount of memory we are willing to waste,
+**| since a block request barely less than this can cause the current hunk
+**| to be retired (with any unused space wasted) as well get a new hunk.
+|*/
+#define morkZone_kMaxHunkWaste ((mork_size) 4096) /* 1/16 kNewHunkSize */
+
+/*| kRound*: the algorithm for rounding up allocation sizes for caching
+**| in free lists works like the following.  We add kRoundAdd to any size
+**| requested, and then bitwise AND with kRoundMask, and this will give us
+**| the smallest multiple of kRoundSize that is at least as large as the
+**| requested size.  Then if we rightshift this number by kRoundBits, we
+**| will have the index into the mZone_FreeRuns array which will hold any
+**| cache runs of that size.  So 4 bits of shift gives us a granularity
+**| of 16 bytes, so that free lists will hold successive runs that are
+**| 16 bytes greater than the next smaller run size.  If we have 256 free
+**| lists of nonzero sized runs altogether, then the largest run that can
+**| be cached is 4096, or 4K (since 4096 == 16 * 256).  A larger run that
+**| gets freed will go in to the free hunk list (or back to the heap).
+|*/
+#define morkZone_kRoundBits 4 /* bits to round-up size for free lists */
+#define morkZone_kRoundSize (1 << morkZone_kRoundBits)
+#define morkZone_kRoundAdd ((1 << morkZone_kRoundBits) - 1)
+#define morkZone_kRoundMask (~ ((mork_ip) morkZone_kRoundAdd))
+
+#define morkZone_kBuckets 256 /* number of distinct free lists */
+
+/*| kMaxCachedRun: the largest run that will be stored inside a free
+**| list of old zapped runs.  A run larger than this cannot be put in
+**| a free list, and must be allocated from the heap at need, and put
+**| into the free hunk list when discarded.
+|*/
+#define morkZone_kMaxCachedRun (morkZone_kBuckets * morkZone_kRoundSize)
+
+#define morkDerived_kZone     /*i*/ 0x5A6E /* ascii 'Zn' */
+
+/*| morkZone: a pooled memory allocator like an NSPR arena.  The term 'zone'
+**| is roughly synonymous with 'heap'.  I avoid calling this class a "heap"
+**| to avoid any confusion with nsIMdbHeap, and I avoid calling this class
+**| an arean to avoid confusion with NSPR usage.
+|*/
+class morkZone : public morkNode, public nsIMdbHeap {
+
+// public: // slots inherited from morkNode (meant to inform only)
+  // nsIMdbHeap*       mNode_Heap;
+
+  // mork_base      mNode_Base;     // must equal morkBase_kNode
+  // mork_derived   mNode_Derived;  // depends on specific node subclass
+  
+  // mork_access    mNode_Access;   // kOpen, kClosing, kShut, or kDead
+  // mork_usage     mNode_Usage;    // kHeap, kStack, kMember, kGlobal, kNone
+  // mork_able      mNode_Mutable;  // can this node be modified?
+  // mork_load      mNode_Load;     // is this node clean or dirty?
+  
+  // mork_uses      mNode_Uses;     // refcount for strong refs
+  // mork_refs      mNode_Refs;     // refcount for strong refs + weak refs
+
+public: // state is public because the entire Mork system is private
+
+  nsIMdbHeap*  mZone_Heap; // strong ref to heap allocating all space
+  
+  mork_size    mZone_HeapVolume;  // total bytes allocated from heap
+  mork_size    mZone_BlockVolume; // total bytes in all zone blocks
+  mork_size    mZone_RunVolume;   // total bytes in all zone runs
+  mork_size    mZone_ChipVolume;  // total bytes in all zone chips
+  
+  mork_size    mZone_FreeOldRunVolume; // total bytes in all used hunks
+  
+  mork_count   mZone_HunkCount;        // total number of used hunks
+  mork_count   mZone_FreeOldRunCount;  // total free old runs
+
+  morkHunk*    mZone_HunkList;       // linked list of all used hunks  
+  morkRun*     mZone_FreeOldRunList; // linked list of free old runs
+  
+  // note mZone_At is a byte pointer for single byte address arithmetic:
+  mork_u1*     mZone_At;     // current position in most recent hunk
+  mork_size    mZone_AtSize; // number of bytes remaining in this hunk
+  
+  // kBuckets+1 so indexes zero through kBuckets are all okay to use:
+  
+  morkRun*     mZone_FreeRuns[ morkZone_kBuckets + 1 ];
+  // Each piece of memory stored in list mZone_FreeRuns[ i ] has an
+  // allocation size equal to sizeof(morkRun) + (i * kRoundSize), so
+  // that callers can be given a piece of memory with (i * kRoundSize)
+  // bytes of writeable space while reserving the first sizeof(morkRun)
+  // bytes to keep track of size information for later re-use.  Note
+  // that mZone_FreeRuns[ 0 ] is unused because no run will be zero
+  // bytes in size (and morkZone plans to complain about zero sizes). 
+
+protected: // zone utilities
+  
+  mork_size zone_grow_at(morkEnv* ev, mork_size inNeededSize);
+
+  void*     zone_new_chip(morkEnv* ev, mdb_size inSize); // alloc  
+  morkHunk* zone_new_hunk(morkEnv* ev, mdb_size inRunSize); // alloc  
+
+// { ===== begin nsIMdbHeap methods =====
+public:
+  NS_IMETHOD Alloc(nsIMdbEnv* ev, // allocate a piece of memory
+    mdb_size inSize,   // requested size of new memory block 
+    void** outBlock) override; // memory block of inSize bytes, or nil
+    
+  NS_IMETHOD Free(nsIMdbEnv* ev, // free block allocated earlier by Alloc()
+    void* inBlock) override;
+
+  virtual size_t GetUsedSize() override { return mZone_Heap->GetUsedSize(); }
+// } ===== end nsIMdbHeap methods =====
+
+// { ===== begin morkNode interface =====
+public: // morkNode virtual methods
+  virtual void CloseMorkNode(morkEnv* ev) override; // CloseZone() only if open
+  virtual ~morkZone(); // assert that CloseMap() executed earlier
+
+public: // morkMap construction & destruction
+  morkZone(morkEnv* ev, const morkUsage& inUsage, nsIMdbHeap* ioNodeHeap, 
+    nsIMdbHeap* ioZoneHeap);
+  
+  void CloseZone(morkEnv* ev); // called by CloseMorkNode()
+  
+public: // dynamic type identification
+  mork_bool IsZone() const
+  { return IsNode() && mNode_Derived == morkDerived_kZone; }
+// } ===== end morkNode methods =====
+
+// { ===== begin morkZone methods =====
+public: // chips do not know how big they are...
+  void* ZoneNewChip(morkEnv* ev, mdb_size inSize); // alloc  
+    
+public: // ...but runs do indeed know how big they are
+  void* ZoneNewRun(morkEnv* ev, mdb_size inSize); // alloc  
+  void  ZoneZapRun(morkEnv* ev, void* ioRunBody); // free
+  void* ZoneGrowRun(morkEnv* ev, void* ioRunBody, mdb_size inSize); // realloc
+    
+// } ===== end morkZone methods =====
+
+public: // typing & errors
+  static void NonZoneTypeError(morkEnv* ev);
+  static void NilZoneHeapError(morkEnv* ev);
+  static void BadZoneTagError(morkEnv* ev);
+};
+
+//3456789_123456789_123456789_123456789_123456789_123456789_123456789_123456789
+
+#endif /* _MORKZONE_ */