
/*--------------------------------------------------------------------*/
/*--- Helgrind: a Valgrind tool for detecting errors               ---*/
/*--- in threaded programs.                              hg_main.c ---*/
/*--------------------------------------------------------------------*/

/*
   This file is part of Helgrind, a Valgrind tool for detecting errors
   in threaded programs.

   Copyright (C) 2007-2007 OpenWorks LLP
      info@open-works.co.uk

   This program is free software; you can redistribute it and/or
   modify it under the terms of the GNU General Public License as
   published by the Free Software Foundation; either version 2 of the
   License, or (at your option) any later version.

   This program is distributed in the hope that it will be useful, but
   WITHOUT ANY WARRANTY; without even the implied warranty of
   MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the GNU
   General Public License for more details.

   You should have received a copy of the GNU General Public License
   along with this program; if not, write to the Free Software
   Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA
   02111-1307, USA.

   The GNU General Public License is contained in the file COPYING.

   Neither the names of the U.S. Department of Energy nor the
   University of California nor the names of its contributors may be
   used to endorse or promote products derived from this software
   without prior written permission.
*/

#include "pub_tool_basics.h"
#include "pub_tool_aspacemgr.h"
#include "pub_tool_libcassert.h"
#include "pub_tool_libcbase.h"
#include "pub_tool_libcprint.h"
#include "pub_tool_mallocfree.h"
#include "pub_tool_threadstate.h"
#include "pub_tool_tooliface.h"
#include "pub_tool_hashtable.h"
#include "pub_tool_replacemalloc.h"
#include "pub_tool_machine.h"
#include "pub_tool_options.h"
#include "pub_tool_xarray.h"
#include "pub_tool_stacktrace.h"

#include "helgrind.h"

#define HG_(str) VGAPPEND(vgHelgrind_,str)
#include "hg_wordfm.h"
#include "hg_wordset.h"

/*----------------------------------------------------------------*/
/*---                                                          ---*/
/*----------------------------------------------------------------*/

/* Note there are a whole bunch of ugly double casts of the form
   (Word*)(void*)&p.  These placate gcc at -O2.  The obvious form
   (Word*)&p causes gcc to complain that 'dereferencing a type-punned
   pointer ill break strict-aliasing rules'.  It stops complaining
   when the intermediate void* type is inserted.  Is this a reasonable
   "fix"?  I don't know. */

// FIXME what is supposed to happen to locks in memory which
// is relocated as a result of client realloc?

// FIXME some kind of ownership recycling problem in
// init_thread_specific_state() for programs which use the same thread
// slot more than once?

// FIXME put referencing ThreadId into Thread and get
// rid of the slow reverse mapping function.

// FIXME accesses to NoAccess areas: change state to Excl?

// FIXME report errors for accesses of NoAccess memory?

// FIXME pth_cond_wait/timedwait wrappers.  Even if these fail,
// the thread still holds the lock.

/* ------------ Debug/trace options ------------ */

// this is:
// shadow_mem_make_NoAccess: 29156 SMs, 1728 scanned
// happens_before_wrk: 1000
// ev__post_thread_join: 3360 SMs, 29 scanned, 252 re-Excls
#define SHOW_EXPENSIVE_STUFF 0

// 0 for silent, 1 for some stuff, 2 for lots of stuff
#define SHOW_EVENTS 0

// Flags for controlling for which events sanity checking is done
#define SCE_THREADS  (1<<0)  // Sanity check at thread create/join
#define SCE_LOCKS    (1<<1)  // Sanity check at lock events
#define SCE_BIGRANGE (1<<2)  // Sanity check at big mem range events
#define SCE_ACCESS   (1<<3)  // Sanity check at mem accesses
#define SCE_LAOG     (1<<4)  // Sanity check at significant LAOG events

#define SCE_BIGRANGE_T 256  // big mem range minimum size


/* For the shadow mem cache stuff we may want more intrusive
   checks.  Unfortunately there's no almost-zero-cost way to make them
   selectable at run time.  Hence set the #if 0 to #if 1 and
   rebuild if you want them. */
#if 0
#  define SCE_CACHELINE 1  /* do sanity-check CacheLine stuff */
#  define inline __attribute__((noinline))
   /* probably want to ditch -fomit-frame-pointer too */
#else
#  define SCE_CACHELINE 0   /* don't sanity-check CacheLine stuff */
#endif

static void all__sanity_check ( Char* who ); /* fwds */

#define HG_CLI__MALLOC_REDZONE_SZB 16 /* let's say */

// 0 for none, 1 for dump at end of run
#define SHOW_DATA_STRUCTURES 0


/* ------------ Command line options ------------ */

// 0 = no segments at all
// 1 = segments at thread create/join
// 2 = as 1 + segments at condition variable signal/broadcast/wait too
static Int clo_happens_before = 2;  /* default setting */

/* Generate .vcg output of the happens-before graph?
   0: no  1: yes, without VTSs  2: yes, with VTSs */
static Int clo_gen_vcg = 0;

/* When comparing race errors for equality, should the race address be
   taken into account?  For users, no, but for verification purposes
   (regtesting) this is sometimes important. */
static Bool clo_cmp_race_err_addrs = False;

/* Tracing memory accesses, so we can see what's going on.
   clo_trace_addr is the address to monitor.  clo_trace_level = 0 for
   no tracing, 1 for summary, 2 for detailed. */
static Addr clo_trace_addr  = 0;
static Int  clo_trace_level = 0;

/* Sanity check level.  This is an or-ing of
   SCE_{THREADS,LOCKS,BIGRANGE,ACCESS,LAOG}. */
static Int clo_sanity_flags = 0;

/* This has to do with printing error messages.  See comments on
   announce_threadset() and summarise_threadset().  Perhaps it
   should be a command line option. */
#define N_THREADS_TO_ANNOUNCE 5


/* ------------ Misc comments ------------ */

// FIXME: don't hardwire initial entries for root thread.
// Instead, let the pre_thread_ll_create handler do this.

// FIXME: when a SecMap is completely set via and address range
// setting operation to a non-ShR/M state, clear its .mbHasShared 
// bit

/* FIXME: figure out what the real rules are for Excl->ShR/M
   transitions w.r.t locksets.

   Muelenfeld thesis Sec 2.2.1 p 8/9 says that

     When another thread accesses the memory location, the lock-set
     is initialized with all active locks and the algorithm reports
     the next access that results in an empty lock-set.

   What does "all active locks" mean?  All locks held by the accessing
   thread, or all locks held by the system as a whole?

   However: Muelenfeld's enhanced Helgrind (eraser_mem_read_word)
   seems to use simply the set of locks held by the thread causing the
   transition into a shared state at the time of the transition:

     *sword = SW(Vge_Shar, packLockSet(thread_locks_rd[tid]));

   Original Eraser paper also says "all active locks".
*/

// Major stuff to fix:
// - reader-writer locks

/* Thread async exit:
   
   remove the map_threads entry
   leave the Thread object in place
   complain if holds any locks
   
   unlike with Join, do not change any memory states

   I _think_ this is correctly handled now.
*/

/*----------------------------------------------------------------*/
/*--- Some very basic stuff                                    ---*/
/*----------------------------------------------------------------*/

static void* hg_zalloc ( SizeT n ) {
   void* p;
   tl_assert(n > 0);
   p = VG_(malloc)( n );
   tl_assert(p);
   VG_(memset)(p, 0, n);
   return p;
}
static void hg_free ( void* p ) {
   tl_assert(p);
   VG_(free)(p);
}

/* Round a up to the next multiple of N.  N must be a power of 2 */
#define ROUNDUP(a, N)   ((a + N - 1) & ~(N-1))
/* Round a down to the next multiple of N.  N must be a power of 2 */
#define ROUNDDN(a, N)   ((a) & ~(N-1))

#ifdef HAVE_BUILTIN_EXPECT
#define LIKELY(cond)   __builtin_expect((cond),1)
#define UNLIKELY(cond) __builtin_expect((cond),0)
#else
#define LIKELY(cond)   (cond)
#define UNLIKELY(cond) (cond)
#endif


/*----------------------------------------------------------------*/
/*--- Primary data definitions                                 ---*/
/*----------------------------------------------------------------*/

/* These are handles for thread segments.  CONSTRAINTS: Must be small
   ints numbered from zero, since 30-bit versions of them must are
   used to represent Exclusive shadow states.  Are used as keys in
   WordFMs so must be castable to Words at the appropriate points. */
typedef  UInt  SegmentID;


/* These are handles for Word sets.  CONSTRAINTS: must be (very) small
   ints numbered from zero, since < 30-bit versions of them are used to
   encode thread-sets and lock-sets in 32-bit shadow words. */
typedef  WordSet  WordSetID;


/* Stores information about a thread.  Addresses of these also serve
   as unique thread identifiers and so are never freed, so they should
   be as small as possible. */
typedef
   struct _Thread {
      /* ADMIN */
      struct _Thread* admin;
      UInt            magic;
      /* USEFUL */
      WordSetID locksetA; /* WordSet of Lock* currently held by thread */
      WordSetID locksetW; /* subset of locksetA held in w-mode */
      SegmentID csegid;  /* current thread segment for thread */
      /* EXPOSITION */
      /* Place where parent was when this thread was created. */
      ExeContext* created_at;
      Bool        announced;
      /* Index for generating references in error messages. */
      Int         errmsg_index;
   }
   Thread;


/* Stores information about a lock's current state.  These are
   allocated and later freed (when the containing memory becomes
   NoAccess).  This gives a problem for the XError type, which
   contains Lock*s.  Solution is to copy any Lock which is to be
   incorporated into an XErrors, so as to make it independent from the
   'normal' collection of Locks, which can come and go.  When the lock
   is copied, its .magic is changed from LockN_Magic to
   LockP_Magic. */

/* Lock kinds. */
typedef
   enum {
      LK_mbRec=1001, /* normal mutex, possibly recursive */
      LK_nonRec,     /* normal mutex, definitely non recursive */
      LK_rdwr        /* reader-writer lock */
   }
   LockKind;

typedef
   struct _Lock {
      /* ADMIN */
      struct _Lock* admin;
      ULong         unique; /* used for persistence-hashing */
      UInt          magic;  /* LockN_MAGIC or LockP_MAGIC */
      /* EXPOSITION */
      /* Place where lock first came to the attention of Helgrind. */
      ExeContext*   appeared_at;
      /* Place where the lock most recently made an unlocked->locked
         transition. */
      ExeContext*   acquired_at;
      /* USEFUL-STATIC */
      Addr          guestaddr; /* Guest address of lock */
      LockKind      kind;      /* what kind of lock this is */
      /* USEFUL-DYNAMIC */
      Bool          heldW; 
      WordBag*      heldBy; /* bag of threads that hold this lock */
      /* .heldBy is NULL: lock is unheld, and .heldW is meaningless
                          but arbitrarily set to False
         .heldBy is non-NULL:
            .heldW is True:  lock is w-held by threads in heldBy
            .heldW is False: lock is r-held by threads in heldBy
            Either way, heldBy may not validly be an empty Bag.

         for LK_nonRec, r-holdings are not allowed, and w-holdings may
         only have sizeTotal(heldBy) == 1

         for LK_mbRec, r-holdings are not allowed, and w-holdings may
         only have sizeUnique(heldBy) == 1

         for LK_rdwr, w-holdings may only have sizeTotal(heldBy) == 1 */
   }
   Lock;


/* Stores information about thread segments.  .prev can be NULL only
   when this is the first segment for the thread.  .other is NULL
   unless this segment depends on a message (create, join, signal)
   from some other thread.  Segments are never freed (!) */
typedef
   struct _Segment {
      /* ADMIN */
      struct _Segment* admin;
      UInt             magic;
      /* USEFUL */
      UInt             dfsver; /* Version # for depth-first searches */
      Thread*          thr;    /* The thread that I am part of */
      struct _Segment* prev;   /* The previous segment in this thread */
      struct _Segment* other;  /* Possibly a segment from some other 
                                  thread, which happened-before me */
      XArray*          vts;    /* XArray of ScalarTS */
      /* DEBUGGING ONLY: what does 'other' arise from?  
         c=thread creation, j=join, s=cvsignal, S=semaphore */
      Char other_hint;
   }
   Segment;


/* ------ CacheLine ------ */

#define N_LINE_BITS      5 /* must be >= 3 */
#define N_LINE_ARANGE    (1 << N_LINE_BITS)
#define N_LINE_TREES     (N_LINE_ARANGE >> 3)

typedef
   struct {
      UShort descrs[N_LINE_TREES];
      UInt   svals[N_LINE_ARANGE]; // == N_LINE_TREES * 8
   }
   CacheLine;

#define TREE_DESCR_16_0 (1<<0)
#define TREE_DESCR_32_0 (1<<1)
#define TREE_DESCR_16_1 (1<<2)
#define TREE_DESCR_64   (1<<3)
#define TREE_DESCR_16_2 (1<<4)
#define TREE_DESCR_32_1 (1<<5)
#define TREE_DESCR_16_3 (1<<6)
#define TREE_DESCR_8_0  (1<<7)
#define TREE_DESCR_8_1  (1<<8)
#define TREE_DESCR_8_2  (1<<9)
#define TREE_DESCR_8_3  (1<<10)
#define TREE_DESCR_8_4  (1<<11)
#define TREE_DESCR_8_5  (1<<12)
#define TREE_DESCR_8_6  (1<<13)
#define TREE_DESCR_8_7  (1<<14)
#define TREE_DESCR_DTY  (1<<15)

typedef
   struct {
      UInt  dict[4]; /* can represent up to 4 diff values in the line */
      UChar ix2s[N_LINE_ARANGE/4]; /* array of N_LINE_ARANGE 2-bit
                                      dict indexes */
      /* if dict[0] == 0 then dict[1] is the index of the CacheLineF
         to use */
   }
   CacheLineZ; /* compressed rep for a cache line */

typedef
   struct {
      Bool inUse;
      UInt w32s[N_LINE_ARANGE];
   }
   CacheLineF; /* full rep for a cache line */


/* Shadow memory.
   Primary map is a WordFM Addr SecMap*.  
   SecMaps cover some page-size-ish section of address space and hold
     a compressed representation.
   CacheLine-sized chunks of SecMaps are copied into a Cache, being
   decompressed when moved into the cache and recompressed on the
   way out.  Because of this, the cache must operate as a writeback
   cache, not a writethrough one.
*/
/* See comments below on shadow_mem_make_NoAccess re performance
   effects of N_SECMAP_BITS settings.  On a 2.4GHz Core2,
   starting/quitting OOo (32-bit), I have these rough numbers:
      N_SECMAP_BITS = 11    2m23
      N_SECMAP_BITS = 12    1m58
      N_SECMAP_BITS = 13    1m53

   Each SecMap must hold a power-of-2 number of CacheLines.  Hence
   N_SECMAP_BITS must >= N_LINE_BITS.
*/
#define N_SECMAP_BITS   13
#define N_SECMAP_ARANGE (1 << N_SECMAP_BITS)

// # CacheLines held by a SecMap
#define N_SECMAP_ZLINES (N_SECMAP_ARANGE / N_LINE_ARANGE)
typedef
   struct {
      UInt magic;
      Bool mbHasLocks;  /* hint: any locks in range?  safe: True */
      Bool mbHasShared; /* hint: any ShM/ShR states in range?  safe: True */
      CacheLineZ  linesZ[N_SECMAP_ZLINES];
      CacheLineF* linesF;
      Int         linesF_size;
   }
   SecMap;

typedef
   struct {
      Int line_no; /* which Z-line are we in? */
      Int word_no; /* inside the line, which word is current? */
   }
   SecMapIter;

static void initSecMapIter ( SecMapIter* itr ) {
   itr->line_no = 0;
   itr->word_no = 0;
}

/* Get the current val, and move to the next position.  This is called
   a huge amount in some programs (eg OpenOffice).  Hence the
   'inline'. */
static UWord stats__secmap_iterator_steppings; /* fwds */

inline
static Bool stepSecMapIter ( /*OUT*/UInt** pVal, 
                             /*MOD*/SecMapIter* itr, SecMap* sm )
{
   CacheLineZ* lineZ = NULL;
   CacheLineF* lineF = NULL;
   /* Either it points to a valid place, or to (-1,-1) */
   stats__secmap_iterator_steppings++;
   if (UNLIKELY(itr->line_no == -1)) {
      tl_assert(itr->word_no == -1);
      return False;
   }
   /* so now it must be a valid place in the SecMap. */
   if (0) VG_(printf)("%p %d %d\n", sm, (Int)itr->line_no, (Int)itr->word_no);
   tl_assert(itr->line_no >= 0 && itr->line_no < N_SECMAP_ZLINES);
   lineZ = &sm->linesZ[itr->line_no];
   if (UNLIKELY(lineZ->dict[0] == 0)) {
      tl_assert(sm->linesF);
      tl_assert(sm->linesF_size > 0);
      tl_assert(lineZ->dict[1] >= 0);
      tl_assert(lineZ->dict[1] < sm->linesF_size);
      lineF = &sm->linesF[ lineZ->dict[1] ];
      tl_assert(lineF->inUse);
      tl_assert(itr->word_no >= 0 && itr->word_no < N_LINE_ARANGE);
      *pVal = &lineF->w32s[itr->word_no];
      itr->word_no++;
      if (itr->word_no == N_LINE_ARANGE)
         itr->word_no = 0;
   } else {
      tl_assert(itr->word_no >= 0 && itr->word_no <= 3);
      tl_assert(lineZ->dict[itr->word_no] != 0);
      *pVal = &lineZ->dict[itr->word_no];
      itr->word_no++;
      if (itr->word_no == 4 || lineZ->dict[itr->word_no] == 0)
         itr->word_no = 0;
   }

   if (itr->word_no == 0) {
      itr->line_no++;
      if (itr->line_no == N_SECMAP_ZLINES) {
         itr->line_no = -1;
         itr->word_no = -1;
      }
   }

   return True;
}

/* ------ Cache ------ */

#define N_WAY_BITS 16
#define N_WAY_NENT (1 << N_WAY_BITS)

/* Each tag is the address of the associated CacheLine, rounded down
   to a CacheLine address boundary.  A CacheLine size must be a power
   of 2 and must be 8 or more.  Hence an easy way to initialise the
   cache so it is empty is to set all the tag values to any value % 8
   != 0, eg 1.  This means all queries in the cache initially miss.
   It does however require us to detect and not writeback, any line
   with a bogus tag. */
typedef
   struct {
      CacheLine lyns0[N_WAY_NENT];
      Addr      tags0[N_WAY_NENT];
   }
   Cache;


/* --------- Primary data structures --------- */

/* Admin linked list of Threads */
static Thread* admin_threads = NULL;

/* Admin linked list of Locks */
static Lock* admin_locks = NULL;

/* Admin linked list of Segments */
static Segment* admin_segments = NULL;

/* Shadow memory primary map */
static WordFM* map_shmem = NULL; /* WordFM Addr SecMap* */
static Cache   cache_shmem;

/* Mapping table for core ThreadIds to Thread* */
static Thread** map_threads = NULL; /* Array[VG_N_THREADS] of Thread* */

/* Mapping table for thread segments IDs to Segment* */
static WordFM* map_segments = NULL; /* WordFM SegmentID Segment* */

/* Mapping table for lock guest addresses to Lock* */
static WordFM* map_locks = NULL; /* WordFM LockAddr Lock* */

/* The word-set universes for thread sets and lock sets. */
static WordSetU* univ_tsets = NULL; /* sets of Thread* */
static WordSetU* univ_lsets = NULL; /* sets of Lock* */
static WordSetU* univ_laog  = NULL; /* sets of Lock*, for LAOG */

/* never changed; we only care about its address.  Is treated as if it
   was a standard userspace lock.  Also we have a Lock* describing it
   so it can participate in lock sets in the usual way. */
static Int   __bus_lock = 0;
static Lock* __bus_lock_Lock = NULL;


/*----------------------------------------------------------------*/
/*--- Simple helpers for the data structures                   ---*/
/*----------------------------------------------------------------*/

static UWord stats__lockN_acquires = 0;
static UWord stats__lockN_releases = 0;

static ThreadId map_threads_maybe_reverse_lookup_SLOW ( Thread* ); /*fwds*/

#define Thread_MAGIC   0x504fc5e5
#define LockN_MAGIC    0x6545b557 /* normal nonpersistent locks */
#define LockP_MAGIC    0x755b5456 /* persistent (copied) locks */
#define Segment_MAGIC  0x49e94d81
#define SecMap_MAGIC   0x571e58cb

static UWord stats__mk_Segment = 0;

/* --------- Constructors --------- */

static inline Bool is_sane_LockN ( Lock* lock ); /* fwds */

static Thread* mk_Thread ( SegmentID csegid ) {
   static Int indx      = 1;
   Thread* thread       = hg_zalloc( sizeof(Lock) );
   thread->locksetA     = HG_(emptyWS)( univ_lsets );
   thread->locksetW     = HG_(emptyWS)( univ_lsets );
   thread->csegid       = csegid;
   thread->magic        = Thread_MAGIC;
   thread->created_at   = NULL;
   thread->announced    = False;
   thread->errmsg_index = indx++;
   thread->admin        = admin_threads;
   admin_threads        = thread;
   return thread;
}
// Make a new lock which is unlocked (hence ownerless)
static Lock* mk_LockN ( LockKind kind, Addr guestaddr ) {
   static ULong unique = 0;
   Lock* lock             = hg_zalloc( sizeof(Lock) );
   lock->admin            = admin_locks;
   lock->unique           = unique++;
   lock->magic            = LockN_MAGIC;
   lock->appeared_at      = NULL;
   lock->acquired_at      = NULL;
   lock->guestaddr        = guestaddr;
   lock->kind             = kind;
   lock->heldW            = False;
   lock->heldBy           = NULL;
   tl_assert(is_sane_LockN(lock));
   admin_locks            = lock;
   return lock;
}
static Segment* mk_Segment ( Thread* thr, Segment* prev, Segment* other ) {
   Segment* seg    = hg_zalloc( sizeof(Segment) );
   seg->dfsver     = 0;
   seg->thr        = thr;
   seg->prev       = prev;
   seg->other      = other;
   seg->vts        = NULL;
   seg->other_hint = ' ';
   seg->magic      = Segment_MAGIC;
   seg->admin      = admin_segments;
   admin_segments = seg;
   stats__mk_Segment++;
   return seg;
}

static inline Bool is_sane_Segment ( Segment* seg ) {
   return seg != NULL && seg->magic == Segment_MAGIC;
}
static inline Bool is_sane_Thread ( Thread* thr ) {
   return thr != NULL && thr->magic == Thread_MAGIC;
}

static Bool is_sane_Bag_of_Threads ( WordBag* bag )
{
   Thread* thr;
   Word    count;
   HG_(initIterBag)( bag );
   while (HG_(nextIterBag)( bag, (Word*)(void*)&thr, &count )) {
      if (count < 1) return False;
      if (!is_sane_Thread(thr)) return False;
   }
   HG_(doneIterBag)( bag );
   return True;
}
static Bool is_sane_Lock_BASE ( Lock* lock )
{
   if (lock == NULL
       || (lock->magic != LockN_MAGIC && lock->magic != LockP_MAGIC))
      return False;
   switch (lock->kind) { 
      case LK_mbRec: case LK_nonRec: case LK_rdwr: break; 
      default: return False; 
   }
   if (lock->heldBy == NULL) {
      if (lock->acquired_at != NULL) return False;
      /* Unheld.  We arbitrarily require heldW to be False. */
      return !lock->heldW;
   } else {
      if (lock->acquired_at == NULL) return False;
   }

   /* If heldBy is non-NULL, we require it to contain at least one
      thread. */
   if (HG_(isEmptyBag)(lock->heldBy))
      return False;

   /* Lock is either r- or w-held. */
   if (!is_sane_Bag_of_Threads(lock->heldBy)) 
      return False;
   if (lock->heldW) {
      /* Held in write-mode */
      if ((lock->kind == LK_nonRec || lock->kind == LK_rdwr)
          && !HG_(isSingletonTotalBag)(lock->heldBy))
         return False;
   } else {
      /* Held in read-mode */
      if (lock->kind != LK_rdwr) return False;
   }
   return True;
}
static inline Bool is_sane_LockP ( Lock* lock ) {
   return lock != NULL 
          && lock->magic == LockP_MAGIC
          && is_sane_Lock_BASE(lock);
}
static inline Bool is_sane_LockN ( Lock* lock ) {
   return lock != NULL 
          && lock->magic == LockN_MAGIC
          && is_sane_Lock_BASE(lock);
}
static inline Bool is_sane_LockNorP ( Lock* lock ) {
   return is_sane_Lock_BASE(lock);
}

/* Release storage for a Lock.  Also release storage in .heldBy, if
   any. */
static void del_LockN ( Lock* lk ) 
{
   tl_assert(is_sane_LockN(lk));
   if (lk->heldBy)
      HG_(deleteBag)( lk->heldBy );
   VG_(memset)(lk, 0xAA, sizeof(*lk));
   hg_free(lk);
}

/* Update 'lk' to reflect that 'thr' now has a write-acquisition of
   it.  This is done strictly: only combinations resulting from
   correct program and libpthread behaviour are allowed. */
static void lockN_acquire_writer ( Lock* lk, Thread* thr ) 
{
   tl_assert(is_sane_LockN(lk));
   tl_assert(is_sane_Thread(thr));

   stats__lockN_acquires++;

   /* EXPOSITION only */
   /* We need to keep recording snapshots of where the lock was
      acquired, so as to produce better lock-order error messages. */
   if (lk->acquired_at == NULL) {
      ThreadId tid;
      tl_assert(lk->heldBy == NULL);
      tid = map_threads_maybe_reverse_lookup_SLOW(thr);
      lk->acquired_at
         = VG_(record_ExeContext(tid, 0/*first_ip_delta*/));
   } else {
      tl_assert(lk->heldBy != NULL);
   }
   /* end EXPOSITION only */

   switch (lk->kind) {
      case LK_nonRec:
      case_LK_nonRec:
         tl_assert(lk->heldBy == NULL); /* can't w-lock recursively */
         tl_assert(!lk->heldW);
         lk->heldW  = True;
         lk->heldBy = HG_(newBag)( hg_zalloc, hg_free );
         HG_(addToBag)( lk->heldBy, (Word)thr );
         break;
      case LK_mbRec:
         if (lk->heldBy == NULL)
            goto case_LK_nonRec;
         /* 2nd and subsequent locking of a lock by its owner */
         tl_assert(lk->heldW);
         /* assert: lk is only held by one thread .. */
         tl_assert(HG_(sizeUniqueBag(lk->heldBy)) == 1);
         /* assert: .. and that thread is 'thr'. */
         tl_assert(HG_(elemBag)(lk->heldBy, (Word)thr)
                   == HG_(sizeTotalBag)(lk->heldBy));
         HG_(addToBag)(lk->heldBy, (Word)thr);
         break;
      case LK_rdwr:
         tl_assert(lk->heldBy == NULL && !lk->heldW); /* must be unheld */
         goto case_LK_nonRec;
      default: 
         tl_assert(0);
  }
  tl_assert(is_sane_LockN(lk));
}

static void lockN_acquire_reader ( Lock* lk, Thread* thr )
{
   tl_assert(is_sane_LockN(lk));
   tl_assert(is_sane_Thread(thr));
   /* can only add reader to a reader-writer lock. */
   tl_assert(lk->kind == LK_rdwr);
   /* lk must be free or already r-held. */
   tl_assert(lk->heldBy == NULL 
             || (lk->heldBy != NULL && !lk->heldW));

   stats__lockN_acquires++;

   /* EXPOSITION only */
   /* We need to keep recording snapshots of where the lock was
      acquired, so as to produce better lock-order error messages. */
   if (lk->acquired_at == NULL) {
      ThreadId tid;
      tl_assert(lk->heldBy == NULL);
      tid = map_threads_maybe_reverse_lookup_SLOW(thr);
      lk->acquired_at
         = VG_(record_ExeContext(tid, 0/*first_ip_delta*/));
   } else {
      tl_assert(lk->heldBy != NULL);
   }
   /* end EXPOSITION only */

   if (lk->heldBy) {
      HG_(addToBag)(lk->heldBy, (Word)thr);
   } else {
      lk->heldW  = False;
      lk->heldBy = HG_(newBag)( hg_zalloc, hg_free );
      HG_(addToBag)( lk->heldBy, (Word)thr );
   }
   tl_assert(!lk->heldW);
   tl_assert(is_sane_LockN(lk));
}

/* Update 'lk' to reflect a release of it by 'thr'.  This is done
   strictly: only combinations resulting from correct program and
   libpthread behaviour are allowed. */

static void lockN_release ( Lock* lk, Thread* thr )
{
   Bool b;
   tl_assert(is_sane_LockN(lk));
   tl_assert(is_sane_Thread(thr));
   /* lock must be held by someone */
   tl_assert(lk->heldBy);
   stats__lockN_releases++;
   /* Remove it from the holder set */
   b = HG_(delFromBag)(lk->heldBy, (Word)thr);
   /* thr must actually have been a holder of lk */
   tl_assert(b);
   /* normalise */
   tl_assert(lk->acquired_at);
   if (HG_(isEmptyBag)(lk->heldBy)) {
      HG_(deleteBag)(lk->heldBy);
      lk->heldBy      = NULL;
      lk->heldW       = False;
      lk->acquired_at = NULL;
   }
   tl_assert(is_sane_LockN(lk));
}

static void remove_Lock_from_locksets_of_all_owning_Threads( Lock* lk )
{
   Thread* thr;
   if (!lk->heldBy) {
      tl_assert(!lk->heldW);
      return;
   }
   /* for each thread that holds this lock do ... */
   HG_(initIterBag)( lk->heldBy );
   while (HG_(nextIterBag)( lk->heldBy, (Word*)(void*)&thr, NULL )) {
      tl_assert(is_sane_Thread(thr));
      tl_assert(HG_(elemWS)( univ_lsets,
                             thr->locksetA, (Word)lk ));
      thr->locksetA
         = HG_(delFromWS)( univ_lsets, thr->locksetA, (Word)lk );

      if (lk->heldW) {
         tl_assert(HG_(elemWS)( univ_lsets,
                                thr->locksetW, (Word)lk ));
         thr->locksetW
            = HG_(delFromWS)( univ_lsets, thr->locksetW, (Word)lk );
      }
   }
   HG_(doneIterBag)( lk->heldBy );
}

/* --------- xxxID functions --------- */

/* Proposal (for debugging sanity):

   SegmentIDs from 0x1000000 .. 0x1FFFFFF (16777216)

   All other xxxID handles are invalid.
*/
static inline Bool is_sane_SegmentID ( SegmentID tseg ) {
   return tseg >= 0x1000000 && tseg <= 0x1FFFFFF;
}
static inline Bool is_sane_ThreadId ( ThreadId coretid ) {
   return coretid >= 0 && coretid < VG_N_THREADS;
}
static SegmentID alloc_SegmentID ( void ) {
   static SegmentID next = 0x1000000;
   tl_assert(is_sane_SegmentID(next));
   return next++;
}

/* --------- Shadow memory --------- */

static inline Bool is_valid_scache_tag ( Addr tag ) {
   /* a valid tag should be naturally aligned to the start of
      a CacheLine. */
   return 0 == (tag & (N_LINE_ARANGE - 1));
}

static inline Bool is_sane_SecMap ( SecMap* sm ) {
   return sm != NULL && sm->magic == SecMap_MAGIC;
}

/* Shadow value encodings:

   11 WordSetID:TSID_BITS WordSetID:LSID_BITS  ShM  thread-set lock-set
   10 WordSetID:TSID_BITS WordSetID:LSID_BITS  ShR  thread-set lock-set
   01 TSegmentID:30                            Excl thread-segment
   00 0--(20)--0 10 0000 0000                  New
   00 0--(20)--0 01 0000 0000                  NoAccess
   00 0--(20)--0 00 0000 0000                  Invalid

   TSID_BITS + LSID_BITS must equal 30.
   The elements in thread sets are Thread*, casted to Word.
   The elements in lock sets are Lock*, casted to Word.
*/

#define N_LSID_BITS  17
#define N_LSID_MASK  ((1 << (N_LSID_BITS)) - 1)
#define N_LSID_SHIFT 0

#define N_TSID_BITS  (30 - (N_LSID_BITS))
#define N_TSID_MASK  ((1 << (N_TSID_BITS)) - 1)
#define N_TSID_SHIFT (N_LSID_BITS)

static inline Bool is_sane_WordSetID_LSet ( WordSetID wset ) {
   return wset >= 0 && wset <= N_LSID_MASK;
}
static inline Bool is_sane_WordSetID_TSet ( WordSetID wset ) {
   return wset >= 0 && wset <= N_TSID_MASK;
}


__attribute__((noinline))
__attribute__((noreturn))
static void mk_SHVAL_fail ( WordSetID tset, WordSetID lset, HChar* who ) {
   VG_(printf)("\n");
   VG_(printf)("Helgrind: Fatal internal error -- cannot continue.\n");
   VG_(printf)("Helgrind: mk_SHVAL_ShR(tset=%d,lset=%d): FAILED\n",
               (Int)tset, (Int)lset);
   VG_(printf)("Helgrind: max allowed tset=%d, lset=%d\n",
               (Int)N_TSID_MASK, (Int)N_LSID_MASK);
   VG_(printf)("Helgrind: program has too many thread "
              "sets or lock sets to track.\n");
   tl_assert(0);
}

static inline UInt mk_SHVAL_ShM ( WordSetID tset, WordSetID lset ) {
   if (LIKELY(is_sane_WordSetID_TSet(tset) 
              && is_sane_WordSetID_LSet(lset))) {
      return (UInt)( (3<<30) | (tset << N_TSID_SHIFT) 
                             | (lset << N_LSID_SHIFT));
   } else {
      mk_SHVAL_fail(tset, lset, "mk_SHVAL_ShM");
   }
}
static inline UInt mk_SHVAL_ShR ( WordSetID tset, WordSetID lset ) {
   if (LIKELY(is_sane_WordSetID_TSet(tset) 
              && is_sane_WordSetID_LSet(lset))) {
      return (UInt)( (2<<30) | (tset << N_TSID_SHIFT) 
                             | (lset << N_LSID_SHIFT) );
   } else {
      mk_SHVAL_fail(tset, lset, "mk_SHVAL_ShR");
   }
}
static inline UInt mk_SHVAL_Excl ( SegmentID tseg ) {
   tl_assert(is_sane_SegmentID(tseg));
   return (UInt)( (1<<30) | tseg );
}
#define SHVAL_New      ((UInt)(2<<8))
#define SHVAL_NoAccess ((UInt)(1<<8))
#define SHVAL_Invalid  ((UInt)(0<<8))

static inline Bool is_SHVAL_ShM ( UInt w32 ) { 
   return (w32 >> 30) == 3;
}
static inline Bool is_SHVAL_ShR ( UInt w32 ) {
   return (w32 >> 30) == 2;
}
static inline Bool is_SHVAL_Sh ( UInt w32 ) {
   return (w32 >> 31) == 1;
}
static inline Bool is_SHVAL_Excl ( UInt w32 ) {
   return (w32 >> 30) == 1; 
}
static inline Bool is_SHVAL_New ( UInt w32 ) {
   return w32 == SHVAL_New;
}
static inline Bool is_SHVAL_NoAccess ( UInt w32 ) { 
   return w32 == SHVAL_NoAccess;
}
static inline Bool is_SHVAL_valid ( UInt w32 ) {
   return is_SHVAL_Excl(w32) || is_SHVAL_NoAccess(w32)
          || is_SHVAL_Sh(w32) || is_SHVAL_New(w32);
}

static inline SegmentID un_SHVAL_Excl ( UInt w32 ) {
   tl_assert(is_SHVAL_Excl(w32));
   return w32 & ~(3<<30);
}
static inline WordSetID un_SHVAL_ShR_tset ( UInt w32 ) {
   tl_assert(is_SHVAL_ShR(w32));
   return (w32 >> N_TSID_SHIFT) & N_TSID_MASK;
}
static inline WordSetID un_SHVAL_ShR_lset ( UInt w32 ) {
   tl_assert(is_SHVAL_ShR(w32));
   return (w32 >> N_LSID_SHIFT) & N_LSID_MASK;
}
static inline WordSetID un_SHVAL_ShM_tset ( UInt w32 ) {
   tl_assert(is_SHVAL_ShM(w32));
   return (w32 >> N_TSID_SHIFT) & N_TSID_MASK;
}
static inline WordSetID un_SHVAL_ShM_lset ( UInt w32 ) {
   tl_assert(is_SHVAL_ShM(w32));
   return (w32 >> N_LSID_SHIFT) & N_LSID_MASK;
}
static inline WordSetID un_SHVAL_Sh_tset ( UInt w32 ) {
   tl_assert(is_SHVAL_Sh(w32));
   return (w32 >> N_TSID_SHIFT) & N_TSID_MASK;
}
static inline WordSetID un_SHVAL_Sh_lset ( UInt w32 ) {
   tl_assert(is_SHVAL_Sh(w32));
   return (w32 >> N_LSID_SHIFT) & N_LSID_MASK;
}


/*----------------------------------------------------------------*/
/*--- Print out the primary data structures                    ---*/
/*----------------------------------------------------------------*/

static WordSetID del_BHL ( WordSetID lockset ); /* fwds */
static 
void get_ZF_by_index ( /*OUT*/CacheLineZ** zp, /*OUT*/CacheLineF** fp,
                       SecMap* sm, Int zix ); /* fwds */
static 
Segment* map_segments_maybe_lookup ( SegmentID segid ); /* fwds */

#define PP_THREADS      (1<<1)
#define PP_LOCKS        (1<<2)
#define PP_SEGMENTS     (1<<3)
#define PP_SHMEM_SHARED (1<<4)
#define PP_ALL (PP_THREADS | PP_LOCKS | PP_SEGMENTS | PP_SHMEM_SHARED)


static const Int sHOW_ADMIN = 0;

static void space ( Int n )
{
   Int  i;
   Char spaces[128+1];
   tl_assert(n >= 0 && n < 128);
   if (n == 0)
      return;
   for (i = 0; i < n; i++)
      spaces[i] = ' ';
   spaces[i] = 0;
   tl_assert(i < 128+1);
   VG_(printf)("%s", spaces);
}

static void pp_Thread ( Int d, Thread* t )
{
   space(d+0); VG_(printf)("Thread %p {\n", t);
   if (sHOW_ADMIN) {
   space(d+3); VG_(printf)("admin    %p\n",   t->admin);
   space(d+3); VG_(printf)("magic    0x%x\n", (UInt)t->magic);
   }
   space(d+3); VG_(printf)("locksetA %d\n",   (Int)t->locksetA);
   space(d+3); VG_(printf)("locksetW %d\n",   (Int)t->locksetW);
   space(d+3); VG_(printf)("csegid   0x%x\n", (UInt)t->csegid);
   space(d+0); VG_(printf)("}\n");
}

static void pp_admin_threads ( Int d )
{
   Int     i, n;
   Thread* t;
   for (n = 0, t = admin_threads;  t;  n++, t = t->admin) {
      /* nothing */
   }
   space(d); VG_(printf)("admin_threads (%d records) {\n", n);
   for (i = 0, t = admin_threads;  t;  i++, t = t->admin) {
      if (0) {
         space(n); 
         VG_(printf)("admin_threads record %d of %d:\n", i, n);
      }
      pp_Thread(d+3, t);
   }
   space(d); VG_(printf)("}\n", n);
}

static void pp_map_threads ( Int d )
{
   Int i, n;
   n = 0;
   space(d); VG_(printf)("map_threads ");
   n = 0;
   for (i = 0; i < VG_N_THREADS; i++) {
      if (map_threads[i] != NULL)
         n++;
   }
   VG_(printf)("(%d entries) {\n", n);
   for (i = 0; i < VG_N_THREADS; i++) {
      if (map_threads[i] == NULL)
         continue;
      space(d+3);
      VG_(printf)("coretid %d -> Thread %p\n", i, map_threads[i]);
   }
   space(d); VG_(printf)("}\n");
}

static const HChar* show_LockKind ( LockKind lkk ) {
   switch (lkk) {
      case LK_mbRec:  return "mbRec";
      case LK_nonRec: return "nonRec";
      case LK_rdwr:   return "rdwr";
      default:        tl_assert(0);
   }
}

static void pp_Lock ( Int d, Lock* lk )
{
   space(d+0); VG_(printf)("Lock %p (ga %p) {\n", lk, lk->guestaddr);
   if (sHOW_ADMIN) {
      space(d+3); VG_(printf)("admin  %p\n",   lk->admin);
      space(d+3); VG_(printf)("magic  0x%x\n", (UInt)lk->magic);
   }
   space(d+3); VG_(printf)("unique %llu\n", lk->unique);
   space(d+3); VG_(printf)("kind   %s\n", show_LockKind(lk->kind));
   space(d+3); VG_(printf)("heldW  %s\n", lk->heldW ? "yes" : "no");
   space(d+3); VG_(printf)("heldBy %p", lk->heldBy);
   if (lk->heldBy) {
      Thread* thr;
      Word    count;
      VG_(printf)(" { ");
      HG_(initIterBag)( lk->heldBy );
      while (HG_(nextIterBag)( lk->heldBy, (Word*)(void*)&thr, &count ))
         VG_(printf)("%lu:%p ", count, thr);
      HG_(doneIterBag)( lk->heldBy );
      VG_(printf)("}");
   }
   VG_(printf)("\n");
   space(d+0); VG_(printf)("}\n");
}

static void pp_admin_locks ( Int d )
{
   Int   i, n;
   Lock* lk;
   for (n = 0, lk = admin_locks;  lk;  n++, lk = lk->admin) {
      /* nothing */
   }
   space(d); VG_(printf)("admin_locks (%d records) {\n", n);
   for (i = 0, lk = admin_locks;  lk;  i++, lk = lk->admin) {
      if (0) {
         space(n); 
         VG_(printf)("admin_locks record %d of %d:\n", i, n);
      }
      pp_Lock(d+3, lk);
   }
   space(d); VG_(printf)("}\n", n);
}

static void pp_map_locks ( Int d )
{
   void* gla;
   Lock* lk;
   space(d); VG_(printf)("map_locks (%d entries) {\n",
                         (Int)HG_(sizeFM)( map_locks ));
   HG_(initIterFM)( map_locks );
   while (HG_(nextIterFM)( map_locks, (Word*)(void*)&gla,
                                      (Word*)(void*)&lk )) {
      space(d+3);
      VG_(printf)("guest %p -> Lock %p\n", gla, lk);
   }
   HG_(doneIterFM)( map_locks );
   space(d); VG_(printf)("}\n");
}

static void pp_Segment ( Int d, Segment* s )
{
   space(d+0); VG_(printf)("Segment %p {\n", s);
   if (sHOW_ADMIN) {
   space(d+3); VG_(printf)("admin  %p\n",   s->admin);
   space(d+3); VG_(printf)("magic  0x%x\n", (UInt)s->magic);
   }
   space(d+3); VG_(printf)("dfsver    %u\n", s->dfsver);
   space(d+3); VG_(printf)("thr       %p\n", s->thr);
   space(d+3); VG_(printf)("prev      %p\n", s->prev);
   space(d+3); VG_(printf)("other[%c] %p\n", s->other_hint, s->other);
   space(d+0); VG_(printf)("}\n");
}

static void pp_admin_segments ( Int d )
{
   Int      i, n;
   Segment* s;
   for (n = 0, s = admin_segments;  s;  n++, s = s->admin) {
      /* nothing */
   }
   space(d); VG_(printf)("admin_segments (%d records) {\n", n);
   for (i = 0, s = admin_segments;  s;  i++, s = s->admin) {
      if (0) {
         space(n); 
         VG_(printf)("admin_segments record %d of %d:\n", i, n);
      }
      pp_Segment(d+3, s);
   }
   space(d); VG_(printf)("}\n", n);
}

static void pp_map_segments ( Int d )
{
   SegmentID segid;
   Segment*  seg;
   space(d); VG_(printf)("map_segments (%d entries) {\n", 
                         (Int)HG_(sizeFM)( map_segments ));
   HG_(initIterFM)( map_segments );
   while (HG_(nextIterFM)( map_segments, (Word*)(void*)&segid,
                                         (Word*)(void*)&seg )) {
      space(d+3);
      VG_(printf)("segid 0x%x -> Segment %p\n", (UInt)segid, seg);
   }
   HG_(doneIterFM)( map_segments );
   space(d); VG_(printf)("}\n");
}

static void show_shadow_w32 ( /*OUT*/Char* buf, Int nBuf, UInt w32 )
{
   tl_assert(nBuf-1 >= 99);
   VG_(memset)(buf, 0, nBuf);
   if (is_SHVAL_ShM(w32)) {
      VG_(sprintf)(buf, "ShM(%u,%u)", 
                   un_SHVAL_ShM_tset(w32), un_SHVAL_ShM_lset(w32));
   }
   else
   if (is_SHVAL_ShR(w32)) {
      VG_(sprintf)(buf, "ShR(%u,%u)", 
                   un_SHVAL_ShR_tset(w32), un_SHVAL_ShR_lset(w32));
   }
   else
   if (is_SHVAL_Excl(w32)) {
      VG_(sprintf)(buf, "Excl(%u)", un_SHVAL_Excl(w32));
   }
   else
   if (is_SHVAL_New(w32)) {
      VG_(sprintf)(buf, "%s", "New");
   }
   else
   if (is_SHVAL_NoAccess(w32)) {
      VG_(sprintf)(buf, "%s", "NoAccess");
   }
   else {
      VG_(sprintf)(buf, "Invalid-shadow-word(%u)", w32);
   }
}

static
void show_shadow_w32_for_user ( /*OUT*/Char* buf, Int nBuf, UInt w32 )
{
   tl_assert(nBuf-1 >= 99);
   VG_(memset)(buf, 0, nBuf);
   if (is_SHVAL_ShM(w32)) {
      WordSetID tset = un_SHVAL_ShM_tset(w32);
      WordSetID lset = del_BHL( un_SHVAL_ShM_lset(w32) );
      VG_(sprintf)(buf, "ShMod(#Tset=%d,#Lset=%d)", 
                   HG_(cardinalityWS)(univ_tsets, tset),
                   HG_(cardinalityWS)(univ_lsets, lset));
   }
   else
   if (is_SHVAL_ShR(w32)) {
      WordSetID tset = un_SHVAL_ShR_tset(w32);
      WordSetID lset = del_BHL( un_SHVAL_ShR_lset(w32) );
      VG_(sprintf)(buf, "ShRO(#Tset=%d,#Lset=%d)", 
                   HG_(cardinalityWS)(univ_tsets, tset),
                   HG_(cardinalityWS)(univ_lsets, lset));
   }
   else
   if (is_SHVAL_Excl(w32)) {
      SegmentID segid  = un_SHVAL_Excl(w32);
      Segment*  mb_seg = map_segments_maybe_lookup(segid);
      if (mb_seg && mb_seg->thr && is_sane_Thread(mb_seg->thr)) {
         VG_(sprintf)(buf, "Exclusive(thr#%d)", mb_seg->thr->errmsg_index);
      } else {
         VG_(sprintf)(buf, "Exclusive(segid=%u)", un_SHVAL_Excl(w32));
      }
   }
   else
   if (is_SHVAL_New(w32)) {
      VG_(sprintf)(buf, "%s", "New");
   }
   else
   if (is_SHVAL_NoAccess(w32)) {
      VG_(sprintf)(buf, "%s", "NoAccess");
   }
   else {
      VG_(sprintf)(buf, "Invalid-shadow-word(%u)", w32);
   }
}

static void pp_SecMap_shared ( Int d, SecMap* sm, Addr ga )
{
   Int  i;
#if 0
   Addr a;
   UInt w32;
   Char buf[100];
#endif
   CacheLineZ* lineZ;
   CacheLineF* lineF;
   space(d+0); VG_(printf)("SecMap %p (ga %p) {\n", sm, (void*)ga);

   for (i = 0; i < N_SECMAP_ZLINES; i++) {
      get_ZF_by_index( &lineZ, &lineF, sm, i );
      space(d+3); VG_(printf)("// pp_SecMap_shared: not implemented\n");
   }

#if 0
   for (i = 0; i < N_SECMAP_ARANGE; i++) {
      w32 = sm->w32s[i];
      a   = ga + 1 * i;
      if (! (is_SHVAL_ShM(w32) || is_SHVAL_ShR(w32)))
         continue;
      space(d+3); VG_(printf)("%p -> 0x%08x ", (void*)a, w32);
      show_shadow_w32(buf, sizeof(buf), w32);
      VG_(printf)("%s\n", buf);
   }
#endif

   space(d+0); VG_(printf)("}\n");
}

static void pp_map_shmem_shared ( Int d )
{
   Addr    ga;
   SecMap* sm;
   space(d); VG_(printf)("map_shmem_ShR_and_ShM_only {\n");
   HG_(initIterFM)( map_shmem );
   while (HG_(nextIterFM)( map_shmem, (Word*)(void*)&ga,
                                      (Word*)(void*)&sm )) {
      pp_SecMap_shared( d+3, sm, ga );
   }
   HG_(doneIterFM) ( map_shmem );
   space(d); VG_(printf)("}\n");
}

static void pp_everything ( Int flags, Char* caller )
{
   Int d = 0;
   VG_(printf)("\n");
   VG_(printf)("All_Data_Structures (caller = \"%s\") {\n", caller);
   if (flags & PP_THREADS) {
      VG_(printf)("\n");
      pp_admin_threads(d+3);
      VG_(printf)("\n");
      pp_map_threads(d+3);
   }
   if (flags & PP_LOCKS) {
      VG_(printf)("\n");
      pp_admin_locks(d+3);
      VG_(printf)("\n");
      pp_map_locks(d+3);
   }
   if (flags & PP_SEGMENTS) {
      VG_(printf)("\n");
      pp_admin_segments(d+3);
      VG_(printf)("\n");
      pp_map_segments(d+3);
   }
   if (flags & PP_SHMEM_SHARED) {
      VG_(printf)("\n");
      pp_map_shmem_shared( d+3 );
   }

   VG_(printf)("\n");
   VG_(printf)("}\n");
   VG_(printf)("\n");
}

#undef SHOW_ADMIN


/*----------------------------------------------------------------*/
/*--- Initialise the primary data structures                   ---*/
/*----------------------------------------------------------------*/

/* fwds */
static void map_segments_add ( SegmentID segid, Segment* seg );
static void shmem__invalidate_scache ( void );
static void hbefore__invalidate_cache ( void );
static void shmem__set_mbHasLocks ( Addr a, Bool b );
static Bool shmem__get_mbHasLocks ( Addr a );
static void shadow_mem_set8 ( Thread* uu_thr_acc, Addr a, UInt svNew );
static XArray* singleton_VTS ( Thread* thr, UWord tym );

static void initialise_data_structures ( void )
{
   SegmentID segid;
   Segment*  seg;
   Thread*   thr;

   /* Get everything initialised and zeroed. */
   tl_assert(admin_threads == NULL);
   tl_assert(admin_locks == NULL);
   tl_assert(admin_segments == NULL);

   tl_assert(sizeof(Addr) == sizeof(Word));
   tl_assert(map_shmem == NULL);
   map_shmem = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxed Word cmp*/);
   tl_assert(map_shmem != NULL);
   shmem__invalidate_scache();

   tl_assert(map_threads == NULL);
   map_threads = hg_zalloc( VG_N_THREADS * sizeof(Thread*) );
   tl_assert(map_threads != NULL);

   /* re <=: < on 64-bit platforms, == on 32-bit ones */
   tl_assert(sizeof(SegmentID) <= sizeof(Word));
   tl_assert(sizeof(Segment*) == sizeof(Word));
   tl_assert(map_segments == NULL);
   map_segments = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxed Word cmp*/);
   tl_assert(map_segments != NULL);
   hbefore__invalidate_cache();

   tl_assert(sizeof(Addr) == sizeof(Word));
   tl_assert(map_locks == NULL);
   map_locks = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxed Word cmp*/);
   tl_assert(map_locks != NULL);

   __bus_lock_Lock = mk_LockN( LK_nonRec, (Addr)&__bus_lock );
   tl_assert(is_sane_LockN(__bus_lock_Lock));
   HG_(addToFM)( map_locks, (Word)&__bus_lock, (Word)__bus_lock_Lock );

   tl_assert(univ_tsets == NULL);
   univ_tsets = HG_(newWordSetU)( hg_zalloc, hg_free, 8/*cacheSize*/ );
   tl_assert(univ_tsets != NULL);

   tl_assert(univ_lsets == NULL);
   univ_lsets = HG_(newWordSetU)( hg_zalloc, hg_free, 8/*cacheSize*/ );
   tl_assert(univ_lsets != NULL);

   tl_assert(univ_laog == NULL);
   univ_laog = HG_(newWordSetU)( hg_zalloc, hg_free, 24/*cacheSize*/ );
   tl_assert(univ_laog != NULL);

   /* Set up entries for the root thread */
   // FIXME: this assumes that the first real ThreadId is 1

   /* a segment for the new thread ... */
   // FIXME: code duplication in ev__post_thread_create
   segid = alloc_SegmentID();
   seg   = mk_Segment( NULL, NULL, NULL );
   map_segments_add( segid, seg );

   /* a Thread for the new thread ... */
   thr = mk_Thread( segid );
   seg->thr = thr;

   /* Give the thread a starting-off vector timestamp. */
   seg->vts = singleton_VTS( seg->thr, 1 );

   /* and bind it in the thread-map table.
      FIXME: assumes root ThreadId == 1. */
   map_threads[1] = thr;

   tl_assert(VG_INVALID_THREADID == 0);

   /* Mark the new bus lock correctly (to stop the sanity checks
      complaining) */
   tl_assert( sizeof(__bus_lock) == 4 );
   shadow_mem_set8( NULL/*unused*/, __bus_lock_Lock->guestaddr, 
                                    mk_SHVAL_Excl(segid) );
   shmem__set_mbHasLocks( __bus_lock_Lock->guestaddr, True );

   all__sanity_check("initialise_data_structures");
}


/*----------------------------------------------------------------*/
/*--- map_threads :: WordFM core-ThreadId Thread*              ---*/
/*----------------------------------------------------------------*/

/* Doesn't assert if the relevant map_threads entry is NULL. */
static Thread* map_threads_maybe_lookup ( ThreadId coretid )
{
   Thread* thr;
   tl_assert( is_sane_ThreadId(coretid) );
   thr = map_threads[coretid];
   return thr;
}

/* Asserts if the relevant map_threads entry is NULL. */
static inline Thread* map_threads_lookup ( ThreadId coretid )
{
   Thread* thr;
   tl_assert( is_sane_ThreadId(coretid) );
   thr = map_threads[coretid];
   tl_assert(thr);
   return thr;
}

/* Do a reverse lookup.  Warning: POTENTIALLY SLOW.  Does not assert
   if 'thr' is not found in map_threads. */
static ThreadId map_threads_maybe_reverse_lookup_SLOW ( Thread* thr )
{
   Int i;
   tl_assert(is_sane_Thread(thr));
   /* Check nobody used the invalid-threadid slot */
   tl_assert(VG_INVALID_THREADID >= 0 && VG_INVALID_THREADID < VG_N_THREADS);
   tl_assert(map_threads[VG_INVALID_THREADID] == NULL);
   for (i = 0; i < VG_N_THREADS; i++) {
      if (i != VG_INVALID_THREADID && map_threads[i] == thr)
         return (ThreadId)i;
   }
   return VG_INVALID_THREADID;
}

/* Do a reverse lookup.  Warning: POTENTIALLY SLOW.  Asserts if 'thr'
   is not found in map_threads. */
static ThreadId map_threads_reverse_lookup_SLOW ( Thread* thr )
{
   ThreadId tid = map_threads_maybe_reverse_lookup_SLOW( thr );
   tl_assert(tid != VG_INVALID_THREADID);
   return tid;
}

static void map_threads_delete ( ThreadId coretid )
{
   Thread* thr;
   tl_assert(coretid != 0);
   tl_assert( is_sane_ThreadId(coretid) );
   thr = map_threads[coretid];
   tl_assert(thr);
   map_threads[coretid] = NULL;
}


/*----------------------------------------------------------------*/
/*--- map_locks :: WordFM guest-Addr-of-lock Lock*             ---*/
/*----------------------------------------------------------------*/

/* Make sure there is a lock table entry for the given (lock) guest
   address.  If not, create one of the stated 'kind' in unheld state.
   In any case, return the address of the existing or new Lock. */
static 
Lock* map_locks_lookup_or_create ( LockKind lkk, Addr ga, ThreadId tid )
{
   Bool  found;
   Lock* oldlock = NULL;
   tl_assert(is_sane_ThreadId(tid));
   found = HG_(lookupFM)( map_locks, 
                          NULL, (Word*)(void*)&oldlock, (Word)ga );
   if (!found) {
      Lock* lock = mk_LockN(lkk, ga);
      lock->appeared_at = VG_(record_ExeContext)( tid, 0 );
      tl_assert(is_sane_LockN(lock));
      HG_(addToFM)( map_locks, (Word)ga, (Word)lock );
      tl_assert(oldlock == NULL);
      // mark the relevant secondary map has .mbHasLocks
      shmem__set_mbHasLocks( ga, True );
      return lock;
   } else {
      tl_assert(oldlock != NULL);
      tl_assert(is_sane_LockN(oldlock));
      tl_assert(oldlock->guestaddr == ga);
      // check the relevant secondary map has .mbHasLocks?
      tl_assert(shmem__get_mbHasLocks(ga) == True);
      return oldlock;
   }
}

static Lock* map_locks_maybe_lookup ( Addr ga )
{
   Bool  found;
   Lock* lk = NULL;
   found = HG_(lookupFM)( map_locks, NULL, (Word*)(void*)&lk, (Word)ga );
   tl_assert(found  ?  lk != NULL  :  lk == NULL);
   if (found) {
      // check the relevant secondary map has .mbHasLocks?
      tl_assert(shmem__get_mbHasLocks(ga) == True);
   }
   return lk;
}

static void map_locks_delete ( Addr ga )
{
   Addr  ga2 = 0;
   Lock* lk  = NULL;
   HG_(delFromFM)( map_locks,
                   (Word*)(void*)&ga2, (Word*)(void*)&lk, (Word)ga );
   /* delFromFM produces the val which is being deleted, if it is
      found.  So assert it is non-null; that in effect asserts that we
      are deleting a (ga, Lock) pair which actually exists. */
   tl_assert(lk != NULL);
   tl_assert(ga2 == ga);
}


/*----------------------------------------------------------------*/
/*--- map_segments :: WordFM SegmentID Segment*                ---*/
/*--- the DAG of thread segments                               ---*/
/*----------------------------------------------------------------*/

static void segments__generate_vcg ( void ); /* fwds */

/*--------------- SegmentID to Segment* maps ---------------*/

static Segment* map_segments_lookup ( SegmentID segid )
{
   Bool     found;
   Segment* seg = NULL;
   tl_assert( is_sane_SegmentID(segid) );
   found = HG_(lookupFM)( map_segments,
                          NULL, (Word*)(void*)&seg, (Word)segid );
   tl_assert(found);
   tl_assert(seg != NULL);
   return seg;
}

static Segment* map_segments_maybe_lookup ( SegmentID segid )
{
   Bool     found;
   Segment* seg = NULL;
   tl_assert( is_sane_SegmentID(segid) );
   found = HG_(lookupFM)( map_segments,
                          NULL, (Word*)(void*)&seg, (Word)segid );
   if (!found) tl_assert(seg == NULL);
   return seg;
}

static void map_segments_add ( SegmentID segid, Segment* seg )
{
   /* This is a bit inefficient.  Oh well. */
   tl_assert( !HG_(lookupFM)( map_segments, NULL, NULL, segid ));
   HG_(addToFM)( map_segments, (Word)segid, (Word)seg );
}

/*--------------- to do with Vector Timestamps ---------------*/

/* Scalar Timestamp */
typedef
   struct {
      Thread* thr;
      UWord   tym;
   }
   ScalarTS;

/* Vector Timestamp = XArray* ScalarTS */

static Bool is_sane_VTS ( XArray* vts )
{
   UWord     i, n;
   ScalarTS  *st1, *st2;
   n = VG_(sizeXA)( vts );
   if (n >= 2) {
      for (i = 0; i < n-1; i++) {
         st1 = VG_(indexXA)( vts, i );
         st2 = VG_(indexXA)( vts, i+1 );
         if (st1->thr >= st2->thr)
            return False;
         if (st1->tym == 0 || st2->tym == 0)
            return False;
      }
   }
   return True;
}

static XArray* new_VTS ( void ) {
   return VG_(newXA)( hg_zalloc, hg_free, sizeof(ScalarTS) );
}
static XArray* singleton_VTS ( Thread* thr, UWord tym ) {
   ScalarTS st;
   XArray*  vts;
   tl_assert(thr);
   tl_assert(tym >= 1);
   vts = new_VTS();
   tl_assert(vts);
   st.thr = thr;
   st.tym = tym;
   VG_(addToXA)( vts, &st );
   return vts;
}


static Bool cmpGEQ_VTS ( XArray* a, XArray* b )
{
   Word     ia, ib, useda, usedb;
   UWord    tyma, tymb;
   Thread*  thr;
   ScalarTS *tmpa, *tmpb;

   Bool all_leq = True;
   Bool all_geq = True;

   tl_assert(a);
   tl_assert(b);
   useda = VG_(sizeXA)( a );
   usedb = VG_(sizeXA)( b );

   ia = ib = 0;

   while (1) {

      /* This logic is to enumerate triples (thr, tyma, tymb) drawn
         from a and b in order, where thr is the next Thread*
         occurring in either a or b, and tyma/b are the relevant
         scalar timestamps, taking into account implicit zeroes. */
      tl_assert(ia >= 0 && ia <= useda);
      tl_assert(ib >= 0 && ib <= usedb);
      tmpa = tmpb = NULL;

      if (ia == useda && ib == usedb) {
         /* both empty - done */
         break;
      }
      else
      if (ia == useda && ib != usedb) {
         /* a empty, use up b */
         tmpb = VG_(indexXA)( b, ib );
         thr  = tmpb->thr;
         tyma = 0;
         tymb = tmpb->tym;
         ib++;
      }
      else
      if (ia != useda && ib == usedb) {
         /* b empty, use up a */
         tmpa = VG_(indexXA)( a, ia );
         thr  = tmpa->thr;
         tyma = tmpa->tym;
         tymb = 0;
         ia++;
      }
      else {
         /* both not empty; extract lowest-Thread*'d triple */
         tmpa = VG_(indexXA)( a, ia );
         tmpb = VG_(indexXA)( b, ib );
         if (tmpa->thr < tmpb->thr) {
            /* a has the lowest unconsidered Thread* */
            thr  = tmpa->thr;
            tyma = tmpa->tym;
            tymb = 0;
            ia++;
         }
         else
         if (tmpa->thr > tmpb->thr) {
            /* b has the lowest unconsidered Thread* */
            thr  = tmpb->thr;
            tyma = 0;
            tymb = tmpb->tym;
            ib++;
         } else {
            /* they both next mention the same Thread* */
            tl_assert(tmpa->thr == tmpb->thr);
            thr  = tmpa->thr; /* == tmpb->thr */
            tyma = tmpa->tym;
            tymb = tmpb->tym;
            ia++;
            ib++;
         }
      }

      /* having laboriously determined (thr, tyma, tymb), do something
         useful with it. */
      if (tyma < tymb)
         all_geq = False;
      if (tyma > tymb)
         all_leq = False;
   }

   if (all_leq && all_geq)
      return True; /* PordEQ */
   /* now we know they aren't equal, so either all_leq or all_geq or
      both are false. */
   if (all_leq)
      return False; /* PordLT */
   if (all_geq)
      return True; /* PordGT */
   /* hmm, neither all_geq or all_leq.  This means unordered. */
   return False; /* PordUN */
}


/* Compute max((tick(thra,a),b) into a new XArray.  a and b are
   unchanged.  If neither a nor b supply a value for 'thra',
   assert. */
static
XArray* tickL_and_joinR_VTS ( Thread* thra, XArray* a, XArray* b )
{
   Word     ia, ib, useda, usedb, ticks_found;
   UWord    tyma, tymb, tymMax;
   Thread*  thr;
   XArray*  res;
   ScalarTS *tmpa, *tmpb;

   tl_assert(a);
   tl_assert(b);
   tl_assert(thra);
   useda = VG_(sizeXA)( a );
   usedb = VG_(sizeXA)( b );

   res = new_VTS();
   ia = ib = ticks_found = 0;

   while (1) {

      /* This logic is to enumerate triples (thr, tyma, tymb) drawn
         from a and b in order, where thr is the next Thread*
         occurring in either a or b, and tyma/b are the relevant
         scalar timestamps, taking into account implicit zeroes. */
      tl_assert(ia >= 0 && ia <= useda);
      tl_assert(ib >= 0 && ib <= usedb);
      tmpa = tmpb = NULL;

      if (ia == useda && ib == usedb) {
         /* both empty - done */
         break;
      }
      else
      if (ia == useda && ib != usedb) {
         /* a empty, use up b */
         tmpb = VG_(indexXA)( b, ib );
         thr  = tmpb->thr;
         tyma = 0;
         tymb = tmpb->tym;
         ib++;
      }
      else
      if (ia != useda && ib == usedb) {
         /* b empty, use up a */
         tmpa = VG_(indexXA)( a, ia );
         thr  = tmpa->thr;
         tyma = tmpa->tym;
         tymb = 0;
         ia++;
      }
      else {
         /* both not empty; extract lowest-Thread*'d triple */
         tmpa = VG_(indexXA)( a, ia );
         tmpb = VG_(indexXA)( b, ib );
         if (tmpa->thr < tmpb->thr) {
            /* a has the lowest unconsidered Thread* */
            thr  = tmpa->thr;
            tyma = tmpa->tym;
            tymb = 0;
            ia++;
         }
         else
         if (tmpa->thr > tmpb->thr) {
            /* b has the lowest unconsidered Thread* */
            thr  = tmpb->thr;
            tyma = 0;
            tymb = tmpb->tym;
            ib++;
         } else {
            /* they both next mention the same Thread* */
            tl_assert(tmpa->thr == tmpb->thr);
            thr  = tmpa->thr; /* == tmpb->thr */
            tyma = tmpa->tym;
            tymb = tmpb->tym;
            ia++;
            ib++;
         }
      }

      /* having laboriously determined (thr, tyma, tymb), do something
         useful with it. */
      if (thr == thra) {
         if (tyma > 0) {
            /* VTS 'a' actually supplied this value; it is not a
               default zero.  Do the required 'tick' action. */
            tyma++;
            ticks_found++;
         } else {
            /* 'a' didn't supply this value, so 'b' must have. */
            tl_assert(tymb > 0);
         }
      }
      tymMax = tyma > tymb ? tyma : tymb;
      if (tymMax > 0) {
         ScalarTS st;
         st.thr = thr;
         st.tym = tymMax;
         VG_(addToXA)( res, &st );
      }

   }

   tl_assert(is_sane_VTS( res ));

   if (thra != NULL) {
      tl_assert(ticks_found == 1);
   } else {
      tl_assert(ticks_found == 0);
   }

   return res;
}


/* Do 'vts[me]++', so to speak.  If 'me' does not have an entry in
   'vts', set it to 1 in the returned VTS. */

static XArray* tick_VTS ( Thread* me, XArray* vts ) {
   ScalarTS* here = NULL;
   ScalarTS  tmp;
   XArray*   res;
   Word      i, n; 
   tl_assert(me);
   tl_assert(is_sane_VTS(vts));
   if (0) VG_(printf)("tick vts thrno %ld szin %d\n",
                      (Word)me->errmsg_index, (Int)VG_(sizeXA)(vts) );
   res = new_VTS();
   n = VG_(sizeXA)( vts );
   for (i = 0; i < n; i++) {
      here = VG_(indexXA)( vts, i );
      if (me < here->thr) {
         /* We just went past 'me', without seeing it. */
         tmp.thr = me;
         tmp.tym = 1;
         VG_(addToXA)( res, &tmp );
         tmp = *here;
         VG_(addToXA)( res, &tmp );
         i++;
         break;
      } 
      else if (me == here->thr) {
         tmp = *here;
         tmp.tym++;
         VG_(addToXA)( res, &tmp );
         i++;
         break;
      }
      else /* me > here->thr */ {
         tmp = *here;
         VG_(addToXA)( res, &tmp );
      }
   }
   tl_assert(i >= 0 && i <= n);
   if (i == n && here && here->thr < me) {
      tmp.thr = me;
      tmp.tym = 1;
      VG_(addToXA)( res, &tmp );
   } else {
      for (/*keepgoing*/; i < n; i++) {
         here = VG_(indexXA)( vts, i );
         tmp = *here;
         VG_(addToXA)( res, &tmp );
      }
   }
   tl_assert(is_sane_VTS(res));
   if (0) VG_(printf)("tick vts thrno %ld szou %d\n",
                      (Word)me->errmsg_index, (Int)VG_(sizeXA)(res) );
   return res;
}

static void show_VTS ( HChar* buf, Int nBuf, XArray* vts ) {
   ScalarTS* st;
   HChar     unit[64];
   Word      i, n;
   Int       avail = nBuf;
   tl_assert(avail > 16);
   buf[0] = '[';
   buf[1] = 0;
   n = VG_(sizeXA)( vts );
   for (i = 0; i < n; i++) {
      tl_assert(avail >= 10);
      st = VG_(indexXA)( vts, i );
      VG_(memset)(unit, 0, sizeof(unit));
      VG_(sprintf)(unit, i < n-1 ? "%ld:%ld " : "%ld:%ld",
                         (Word)st->thr->errmsg_index, st->tym);
      if (avail < VG_(strlen)(unit) + 10/*let's say*/) {
         VG_(strcat)(buf, " ...]");
         return;
      }
      VG_(strcat)(buf, unit);
      avail -= VG_(strlen)(unit);
   }
   VG_(strcat)(buf, "]");
}


/*------------ searching the happens-before graph ------------*/

static UWord stats__hbefore_queries   = 0; // total # queries
static UWord stats__hbefore_cache0s   = 0; // hits at cache[0]
static UWord stats__hbefore_cacheNs   = 0; // hits at cache[> 0]
static UWord stats__hbefore_probes    = 0; // # checks in cache
static UWord stats__hbefore_gsearches = 0; // # searches in graph
static UWord stats__hbefore_gsearchFs = 0; // # fast searches in graph
static UWord stats__hbefore_invals    = 0; // # cache invals
static UWord stats__hbefore_stk_hwm   = 0; // stack high water mark

/* Running marker for depth-first searches */
/* NOTE: global variable */
static UInt dfsver_current = 0;

/* A stack of possibly-unexplored nodes used in the depth first search */
/* NOTE: global variable */
static XArray* dfsver_stack = NULL;

// FIXME: check this - is it really correct?
__attribute__((noinline))
static Bool happens_before_do_dfs_from_to ( Segment* src, Segment* dst )
{
   Segment* here;
   Word     ssz;

   /* begin SPEEDUP HACK -- the following can safely be omitted */
   /* fast track common case, without favouring either the
      ->prev or ->other links */
   tl_assert(src);
   tl_assert(dst);
   if ((src->prev && src->prev == dst)
       || (src->other && src->other == dst)) {
      stats__hbefore_gsearchFs++;
      return True;
   }
   /* end SPEEDUP HACK */

   /* empty out the stack */
   tl_assert(dfsver_stack);
   VG_(dropTailXA)( dfsver_stack, VG_(sizeXA)( dfsver_stack ));
   tl_assert(VG_(sizeXA)( dfsver_stack ) == 0);

   /* push starting point */
   (void) VG_(addToXA)( dfsver_stack, &src );

   while (True) {
      /* While the stack is not empty, pop the next node off it and
         consider. */
      ssz = VG_(sizeXA)( dfsver_stack );
      tl_assert(ssz >= 0);
      if (ssz == 0)
         return False; /* stack empty ==> no path from src to dst */

      if (UNLIKELY( ((UWord)ssz) > stats__hbefore_stk_hwm ))
         stats__hbefore_stk_hwm = (UWord)ssz;

      /* here = pop(stack) */
      here = *(Segment**) VG_(indexXA)( dfsver_stack, ssz-1 );
      VG_(dropTailXA)( dfsver_stack, 1 );

     again:
      /* consider the node 'here' */
      if (here == dst)
         return True; /* found a path from src and dst */

      /* have we been to 'here' before? */
      tl_assert(here->dfsver <= dfsver_current);
      if (here->dfsver == dfsver_current)
         continue; /* We've been 'here' before - node is not interesting*/

      /* Mark that we've been here */
      here->dfsver = dfsver_current;

      /* Now push both children on the stack */

      /* begin SPEEDUP hack -- the following can safely be omitted */
      /* idea is, if there is exactly one child, avoid the overhead of
         pushing it on the stack and immediately popping it off again.
         Kinda like doing a tail-call. */
      if (here->prev && !here->other) {
         here = here->prev;
         goto again;
      }
      if (here->other && !here->prev) {
         here = here->other;
         goto again;
      }
      /* end of SPEEDUP HACK */

      /* Push all available children on stack.  From some quick
         experimentation it seems like exploring ->other first leads
         to lower maximum stack use, although getting repeatable
         results is difficult. */
      if (here->prev)
         (void) VG_(addToXA)( dfsver_stack, &(here->prev) );
      if (here->other)
         (void) VG_(addToXA)( dfsver_stack, &(here->other) );
   }
}

__attribute__((noinline))
static Bool happens_before_wrk ( Segment* seg1, Segment* seg2 )
{
   Bool reachable;

   { static Int nnn = 0;
     if (SHOW_EXPENSIVE_STUFF && (nnn++ % 1000) == 0)
        VG_(printf)("happens_before_wrk: %d\n", nnn);
   }

   /* Now the question is, is there a chain of pointers through the
      .prev and .other fields, that leads from seg2 back to seg1 ? */
   tl_assert(dfsver_current < 0xFFFFFFFF);
   dfsver_current++;
   
   if (dfsver_stack == NULL) {
     dfsver_stack = VG_(newXA)( hg_zalloc, hg_free, sizeof(Segment*) );
     tl_assert(dfsver_stack);
   }

   reachable = happens_before_do_dfs_from_to( seg2, seg1 );

   return reachable;
}

/*--------------- the happens_before cache ---------------*/

#define HBEFORE__N_CACHE 64
typedef 
   struct { SegmentID segid1; SegmentID segid2; Bool result; } 
   HBeforeCacheEnt;

static HBeforeCacheEnt hbefore__cache[HBEFORE__N_CACHE];

static void hbefore__invalidate_cache ( void ) 
{
   Int i;
   SegmentID bogus = 0;
   tl_assert(!is_sane_SegmentID(bogus));
   stats__hbefore_invals++;
   for (i = 0; i < HBEFORE__N_CACHE; i++) {
      hbefore__cache[i].segid1 = bogus;
      hbefore__cache[i].segid2 = bogus;
      hbefore__cache[i].result = False;
   }
}

static Bool happens_before ( SegmentID segid1, SegmentID segid2 )
{
   Bool    hbG, hbV;
   Int     i, j, iNSERT_POINT;
   Segment *seg1, *seg2;
   tl_assert(is_sane_SegmentID(segid1));
   tl_assert(is_sane_SegmentID(segid2));
   tl_assert(segid1 != segid2);
   stats__hbefore_queries++;
   stats__hbefore_probes++;
   if (segid1 == hbefore__cache[0].segid1 
       && segid2 == hbefore__cache[0].segid2) {
      stats__hbefore_cache0s++;
      return hbefore__cache[0].result;
   }
   for (i = 1; i < HBEFORE__N_CACHE; i++) {
      stats__hbefore_probes++;
      if (segid1 == hbefore__cache[i].segid1 
          && segid2 == hbefore__cache[i].segid2) {
         /* Found it.  Move it 1 step closer to the front. */
         HBeforeCacheEnt tmp = hbefore__cache[i];
         hbefore__cache[i]   = hbefore__cache[i-1];
         hbefore__cache[i-1] = tmp;
         stats__hbefore_cacheNs++;
         return tmp.result;
      }
   }
   /* Not found.  Search the graph and add an entry to the cache. */
   stats__hbefore_gsearches++;

   seg1 = map_segments_lookup(segid1);
   seg2 = map_segments_lookup(segid2);
   tl_assert(is_sane_Segment(seg1));
   tl_assert(is_sane_Segment(seg2));
   tl_assert(seg1 != seg2);
   tl_assert(seg1->vts);
   tl_assert(seg2->vts);

   hbV = cmpGEQ_VTS( seg2->vts, seg1->vts );
   if (0) {
      /* Crosscheck the vector-timestamp comparison result against that
         obtained from the explicit graph approach.  Can be very
         slow. */
      hbG = happens_before_wrk( seg1, seg2 );
   } else {
      /* Assume the vector-timestamp comparison result is correct, and
         use it as-is. */
      hbG = hbV;
   }

   if (hbV != hbG) {
      VG_(printf)("seg1 %p  seg2 %p  hbV %d  hbG %d\n", 
                  seg1,seg2,(Int)hbV,(Int)hbG);
      segments__generate_vcg();
   }
   tl_assert(hbV == hbG);

   iNSERT_POINT = (1*HBEFORE__N_CACHE)/4 - 1;
   /* if (iNSERT_POINT > 4) iNSERT_POINT = 4; */

   for (j = HBEFORE__N_CACHE-1; j > iNSERT_POINT; j--) {
      hbefore__cache[j] = hbefore__cache[j-1];
   }
   hbefore__cache[iNSERT_POINT].segid1 = segid1;
   hbefore__cache[iNSERT_POINT].segid2 = segid2;
   hbefore__cache[iNSERT_POINT].result = hbG;

   if (0)
   VG_(printf)("hb %d %d\n", (Int)segid1-(1<<24), (Int)segid2-(1<<24));
   return hbG;
}

/*--------------- generating .vcg output ---------------*/

static void segments__generate_vcg ( void )
{
#define PFX "xxxxxx"
   /* Edge colours:
         Black  -- the chain of .prev links
         Green  -- thread creation, link to parent
         Red    -- thread exit, link to exiting thread
         Yellow -- signal edge
         Pink   -- semaphore-up edge
   */
   Segment* seg;
   HChar vtsstr[128];
   VG_(printf)(PFX "graph: { title: \"Segments\"\n");
   VG_(printf)(PFX "orientation: top_to_bottom\n");
   VG_(printf)(PFX "height: 900\n");
   VG_(printf)(PFX "width: 500\n");
   VG_(printf)(PFX "x: 20\n");
   VG_(printf)(PFX "y: 20\n");
   VG_(printf)(PFX "color: lightgrey\n");
   for (seg = admin_segments; seg; seg=seg->admin) {

      VG_(printf)(PFX "node: { title: \"%p\" color: lightcyan "
                  "textcolor: darkgreen label: \"Seg %p\\n", 
                  seg, seg);
      if (seg->thr->errmsg_index == 1) {
         VG_(printf)("ROOT_THREAD");
      } else {
         VG_(printf)("Thr# %d", seg->thr->errmsg_index);
      }

      if (clo_gen_vcg >= 2) {
         show_VTS( vtsstr, sizeof(vtsstr)-1, seg->vts );
         vtsstr[sizeof(vtsstr)-1] = 0;
         VG_(printf)("\\n%s", vtsstr);
      }

      VG_(printf)("\" }\n", vtsstr);

      if (seg->prev)
         VG_(printf)(PFX "edge: { sourcename: \"%p\" targetname: \"%p\""
                     "color: black }\n", seg->prev, seg );
      if (seg->other) {
         HChar* colour = "orange";
         switch (seg->other_hint) {
            case 'c': colour = "darkgreen";  break; /* creation */
            case 'j': colour = "red";        break; /* join (exit) */
            case 's': colour = "orange";     break; /* signal */
            case 'S': colour = "pink";       break; /* sem_post->wait */
            case 'u': colour = "cyan";       break; /* unlock */
            default: tl_assert(0);
         }
         VG_(printf)(PFX "edge: { sourcename: \"%p\" targetname: \"%p\""
                     " color: %s }\n", seg->other, seg, colour );
      }
   }
   VG_(printf)(PFX "}\n");
#undef PFX
}


/*----------------------------------------------------------------*/
/*--- map_shmem :: WordFM Addr SecMap                          ---*/
/*--- shadow memory (low level handlers) (shmem__* fns)        ---*/
/*----------------------------------------------------------------*/


static UWord stats__secmaps_allocd       = 0; // # SecMaps issued
static UWord stats__secmap_ga_space_covered = 0; // # ga bytes covered
static UWord stats__secmap_linesZ_allocd = 0; // # CacheLineZ's issued
static UWord stats__secmap_linesZ_bytes  = 0; // .. using this much storage
static UWord stats__secmap_linesF_allocd = 0; // # CacheLineF's issued
static UWord stats__secmap_linesF_bytes  = 0; //  .. using this much storage
static UWord stats__secmap_iterator_steppings = 0; // # calls to stepSMIter
static UWord stats__cache_Z_fetches      = 0; // # Z lines fetched
static UWord stats__cache_Z_wbacks       = 0; // # Z lines written back
static UWord stats__cache_F_fetches      = 0; // # F lines fetched
static UWord stats__cache_F_wbacks       = 0; // # F lines written back
static UWord stats__cache_invals         = 0; // # cache invals
static UWord stats__cache_flushes        = 0; // # cache flushes
static UWord stats__cache_totrefs        = 0; // # total accesses
static UWord stats__cache_totmisses      = 0; // # misses
static UWord stats__cline_normalises     = 0; // # calls to cacheline_normalise
static UWord stats__cline_read64s        = 0; // # calls to s_m_read64
static UWord stats__cline_read32s        = 0; // # calls to s_m_read32
static UWord stats__cline_read16s        = 0; // # calls to s_m_read16
static UWord stats__cline_read8s         = 0; // # calls to s_m_read8
static UWord stats__cline_write64s       = 0; // # calls to s_m_write64
static UWord stats__cline_write32s       = 0; // # calls to s_m_write32
static UWord stats__cline_write16s       = 0; // # calls to s_m_write16
static UWord stats__cline_write8s        = 0; // # calls to s_m_write8
static UWord stats__cline_set64s         = 0; // # calls to s_m_set64
static UWord stats__cline_set32s         = 0; // # calls to s_m_set32
static UWord stats__cline_set16s         = 0; // # calls to s_m_set16
static UWord stats__cline_set8s          = 0; // # calls to s_m_set8
static UWord stats__cline_get8s          = 0; // # calls to s_m_get8
static UWord stats__cline_copy8s         = 0; // # calls to s_m_copy8
static UWord stats__cline_64to32splits   = 0; // # 64-bit accesses split
static UWord stats__cline_32to16splits   = 0; // # 32-bit accesses split
static UWord stats__cline_16to8splits    = 0; // # 16-bit accesses split
static UWord stats__cline_64to32pulldown = 0; // # calls to pulldown_to_32
static UWord stats__cline_32to16pulldown = 0; // # calls to pulldown_to_16
static UWord stats__cline_16to8pulldown  = 0; // # calls to pulldown_to_8


static UInt shadow_mem_get8 ( Addr a ); /* fwds */

static inline Addr shmem__round_to_SecMap_base ( Addr a ) {
   return a & ~(N_SECMAP_ARANGE - 1);
}
static inline UWord shmem__get_SecMap_offset ( Addr a ) {
   return a & (N_SECMAP_ARANGE - 1);
}

/*--------------- SecMap allocation --------------- */

static HChar* shmem__bigchunk_next = NULL;
static HChar* shmem__bigchunk_end1 = NULL;

static void* shmem__bigchunk_alloc ( SizeT n )
{
   const SizeT sHMEM__BIGCHUNK_SIZE = 4096 * 256;
   tl_assert(n > 0);
   n = ROUNDUP(n, 16);
   tl_assert(shmem__bigchunk_next <= shmem__bigchunk_end1);
   tl_assert(shmem__bigchunk_end1 - shmem__bigchunk_next
             <= (SSizeT)sHMEM__BIGCHUNK_SIZE);
   if (shmem__bigchunk_next + n > shmem__bigchunk_end1) {
      if (0)
      VG_(printf)("XXXXX bigchunk: abandoning %d bytes\n", 
                  (Int)(shmem__bigchunk_end1 - shmem__bigchunk_next));
      shmem__bigchunk_next = VG_(am_shadow_alloc)( sHMEM__BIGCHUNK_SIZE );
      shmem__bigchunk_end1 = shmem__bigchunk_next + sHMEM__BIGCHUNK_SIZE;
   }
   tl_assert(shmem__bigchunk_next);
   tl_assert( 0 == (((Addr)shmem__bigchunk_next) & (16-1)) );
   tl_assert(shmem__bigchunk_next + n <= shmem__bigchunk_end1);
   shmem__bigchunk_next += n;
   return shmem__bigchunk_next - n;
}

static SecMap* shmem__alloc_SecMap ( void )
{
   Word    i, j;
   SecMap* sm = shmem__bigchunk_alloc( sizeof(SecMap) );
   if (0) VG_(printf)("alloc_SecMap %p\n",sm);
   tl_assert(sm);
   sm->magic       = SecMap_MAGIC;
   sm->mbHasLocks  = False; /* dangerous */
   sm->mbHasShared = False; /* dangerous */
   for (i = 0; i < N_SECMAP_ZLINES; i++) {
      sm->linesZ[i].dict[0] = SHVAL_NoAccess;
      sm->linesZ[i].dict[1] = 0; /* completely invalid SHVAL */
      sm->linesZ[i].dict[2] = 0;
      sm->linesZ[i].dict[3] = 0;
      for (j = 0; j < N_LINE_ARANGE/4; j++)
         sm->linesZ[i].ix2s[j] = 0; /* all reference dict[0] */
   }
   sm->linesF      = NULL;
   sm->linesF_size = 0;
   stats__secmaps_allocd++;
   stats__secmap_ga_space_covered += N_SECMAP_ARANGE;
   stats__secmap_linesZ_allocd += N_SECMAP_ZLINES;
   stats__secmap_linesZ_bytes += N_SECMAP_ZLINES * sizeof(CacheLineZ);
   return sm;
}

static SecMap* shmem__find_or_alloc_SecMap ( Addr ga )
{
   SecMap* sm    = NULL;
   Addr    gaKey = shmem__round_to_SecMap_base(ga);
   if (HG_(lookupFM)( map_shmem,
                      NULL/*keyP*/, (Word*)(void*)&sm, (Word)gaKey )) {
      /* Found; address of SecMap is in sm */
      tl_assert(sm);
   } else {
      /* create a new one */
      sm = shmem__alloc_SecMap();
      tl_assert(sm);
      HG_(addToFM)( map_shmem, (Word)gaKey, (Word)sm );
   }
   return sm;
}


/*--------------- cache management/lookup --------------- */

/*--------------- misc --------------- */

static Bool shmem__get_mbHasLocks ( Addr a )
{
   SecMap* sm;
   Addr aKey = shmem__round_to_SecMap_base(a);
   if (HG_(lookupFM)( map_shmem,
                      NULL/*keyP*/, (Word*)(void*)&sm, (Word)aKey )) {
      /* Found */
      return sm->mbHasLocks;
   } else {
      return False;
   }
}

static void shmem__set_mbHasLocks ( Addr a, Bool b )
{
   SecMap* sm;
   Addr aKey = shmem__round_to_SecMap_base(a);
   tl_assert(b == False || b == True);
   if (HG_(lookupFM)( map_shmem,
                      NULL/*keyP*/, (Word*)(void*)&sm, (Word)aKey )) {
      /* Found; address of SecMap is in sm */
   } else {
      /* create a new one */
      sm = shmem__alloc_SecMap();
      tl_assert(sm);
      HG_(addToFM)( map_shmem, (Word)aKey, (Word)sm );
   }
   sm->mbHasLocks = b;
}

static void shmem__set_mbHasShared ( Addr a, Bool b )
{
   SecMap* sm;
   Addr aKey = shmem__round_to_SecMap_base(a);
   tl_assert(b == False || b == True);
   if (HG_(lookupFM)( map_shmem,
                      NULL/*keyP*/, (Word*)(void*)&sm, (Word)aKey )) {
      /* Found; address of SecMap is in sm */
   } else {
      /* create a new one */
      sm = shmem__alloc_SecMap();
      tl_assert(sm);
      HG_(addToFM)( map_shmem, (Word)aKey, (Word)sm );
   }
   sm->mbHasShared = b;
}


/*----------------------------------------------------------------*/
/*--- Sanity checking the data structures                      ---*/
/*----------------------------------------------------------------*/

static UWord stats__sanity_checks = 0;

static Bool is_sane_CacheLine ( CacheLine* cl ); /* fwds */
static Bool cmpGEQ_VTS ( XArray* a, XArray* b ); /* fwds */
static void laog__sanity_check ( Char* who ); /* fwds */

/* REQUIRED INVARIANTS:

   Thread vs Segment/Lock/SecMaps

      for each t in Threads {

         // Thread.lockset: each element is really a valid Lock

         // Thread.lockset: each Lock in set is actually held by that thread
         for lk in Thread.lockset 
            lk == LockedBy(t)

         // Thread.csegid is a valid SegmentID
         // and the associated Segment has .thr == t

      }

      all thread Locksets are pairwise empty under intersection
      (that is, no lock is claimed to be held by more than one thread)
      -- this is guaranteed if all locks in locksets point back to their
      owner threads

   Lock vs Thread/Segment/SecMaps

      for each entry (gla, la) in map_locks
         gla == la->guest_addr

      for each lk in Locks {

         lk->tag is valid
         lk->guest_addr does not have shadow state NoAccess
         if lk == LockedBy(t), then t->lockset contains lk
         if lk == UnlockedBy(segid) then segid is valid SegmentID
             and can be mapped to a valid Segment(seg)
             and seg->thr->lockset does not contain lk
         if lk == UnlockedNew then (no lockset contains lk)

         secmaps for lk has .mbHasLocks == True

      }

   Segment vs Thread/Lock/SecMaps

      the Segment graph is a dag (no cycles)
      all of the Segment graph must be reachable from the segids
         mentioned in the Threads

      for seg in Segments {

         seg->thr is a sane Thread

      }

   SecMaps vs Segment/Thread/Lock

      for sm in SecMaps {

         sm properly aligned
         if any shadow word is ShR or ShM then .mbHasShared == True

         for each Excl(segid) state
            map_segments_lookup maps to a sane Segment(seg)
         for each ShM/ShR(tsetid,lsetid) state
            each lk in lset is a valid Lock
            each thr in tset is a valid thread, which is non-dead

      }
*/


/* Return True iff 'thr' holds 'lk' in some mode. */
static Bool thread_is_a_holder_of_Lock ( Thread* thr, Lock* lk )
{
   if (lk->heldBy)
      return HG_(elemBag)( lk->heldBy, (Word)thr ) > 0;
   else
      return False;
}

/* Sanity check Threads, as far as possible */
__attribute__((noinline))
static void threads__sanity_check ( Char* who )
{
#define BAD(_str) do { how = (_str); goto bad; } while (0)
   Char*     how = "no error";
   Thread*   thr;
   WordSetID wsA, wsW;
   Word*     ls_words;
   Word      ls_size, i;
   Lock*     lk;
   Segment*  seg;
   for (thr = admin_threads; thr; thr = thr->admin) {
      if (!is_sane_Thread(thr)) BAD("1");
      wsA = thr->locksetA;
      wsW = thr->locksetW;
      // locks held in W mode are a subset of all locks held
      if (!HG_(isSubsetOf)( univ_lsets, wsW, wsA )) BAD("7");
      HG_(getPayloadWS)( &ls_words, &ls_size, univ_lsets, wsA );
      for (i = 0; i < ls_size; i++) {
         lk = (Lock*)ls_words[i];
         // Thread.lockset: each element is really a valid Lock
         if (!is_sane_LockN(lk)) BAD("2");
         // Thread.lockset: each Lock in set is actually held by that
         // thread
         if (!thread_is_a_holder_of_Lock(thr,lk)) BAD("3");
         // Thread.csegid is a valid SegmentID
         if (!is_sane_SegmentID(thr->csegid)) BAD("4");
         // and the associated Segment has .thr == t
         seg = map_segments_maybe_lookup(thr->csegid);
         if (!is_sane_Segment(seg)) BAD("5");
         if (seg->thr != thr) BAD("6");
      }
   }
   return;
  bad:
   VG_(printf)("threads__sanity_check: who=\"%s\", bad=\"%s\"\n", who, how);
   tl_assert(0);
#undef BAD
}


/* Sanity check Locks, as far as possible */
__attribute__((noinline))
static void locks__sanity_check ( Char* who )
{
#define BAD(_str) do { how = (_str); goto bad; } while (0)
   Char*     how = "no error";
   Addr      gla;
   Lock*     lk;
   Int       i;
   // # entries in admin_locks == # entries in map_locks
   for (i = 0, lk = admin_locks;  lk;  i++, lk = lk->admin)
      ;
   if (i != HG_(sizeFM)(map_locks)) BAD("1");
   // for each entry (gla, lk) in map_locks
   //      gla == lk->guest_addr
   HG_(initIterFM)( map_locks );
   while (HG_(nextIterFM)( map_locks,
                           (Word*)(void*)&gla, (Word*)(void*)&lk )) {
      if (lk->guestaddr != gla) BAD("2");
   }
   HG_(doneIterFM)( map_locks );
   // scan through admin_locks ...
   for (lk = admin_locks; lk; lk = lk->admin) {
      // lock is sane.  Quite comprehensive, also checks that
      // referenced (holder) threads are sane.
      if (!is_sane_LockN(lk)) BAD("3");
      // map_locks binds guest address back to this lock
      if (lk != map_locks_maybe_lookup(lk->guestaddr)) BAD("4");
      // lk->guest_addr does not have shadow state NoAccess
      // FIXME: this could legitimately arise from a buggy guest
      // that attempts to lock in (eg) freed memory.  Detect this
      // and warn about it in the pre/post-mutex-lock event handler.
      if (is_SHVAL_NoAccess(shadow_mem_get8(lk->guestaddr))) BAD("5");
      // look at all threads mentioned as holders of this lock.  Ensure
      // this lock is mentioned in their locksets.
      if (lk->heldBy) {
         Thread* thr;
         Word    count;
         HG_(initIterBag)( lk->heldBy );
         while (HG_(nextIterBag)( lk->heldBy, 
                                  (Word*)(void*)&thr, &count )) {
            // is_sane_LockN above ensures these
            tl_assert(count >= 1);
            tl_assert(is_sane_Thread(thr));
            if (!HG_(elemWS)(univ_lsets, thr->locksetA, (Word)lk)) 
               BAD("6");
            // also check the w-only lockset
            if (lk->heldW 
                && !HG_(elemWS)(univ_lsets, thr->locksetW, (Word)lk)) 
               BAD("7");
            if ((!lk->heldW)
                && HG_(elemWS)(univ_lsets, thr->locksetW, (Word)lk)) 
               BAD("8");
         }
         HG_(doneIterBag)( lk->heldBy );
      } else {
         /* lock not held by anybody */
         if (lk->heldW) BAD("9"); /* should be False if !heldBy */
         // since lk is unheld, then (no lockset contains lk)
         // hmm, this is really too expensive to check.  Hmm.
      }
      // secmaps for lk has .mbHasLocks == True
      if (!shmem__get_mbHasLocks(lk->guestaddr)) BAD("10");
   }

   return;
  bad:
   VG_(printf)("locks__sanity_check: who=\"%s\", bad=\"%s\"\n", who, how);
   tl_assert(0);
#undef BAD
}


/* Sanity check Segments, as far as possible */
__attribute__((noinline))
static void segments__sanity_check ( Char* who )
{
#define BAD(_str) do { how = (_str); goto bad; } while (0)
   Char*    how = "no error";
   Int      i;
   Segment* seg;
   // FIXME
   //   the Segment graph is a dag (no cycles)
   //   all of the Segment graph must be reachable from the segids
   //      mentioned in the Threads
   // # entries in admin_segments == # entries in map_segments
   for (i = 0, seg = admin_segments;  seg;  i++, seg = seg->admin)
      ;
   if (i != HG_(sizeFM)(map_segments)) BAD("1");
   // for seg in Segments {
   for (seg = admin_segments; seg; seg = seg->admin) {
      if (!is_sane_Segment(seg)) BAD("2");
      if (!is_sane_Thread(seg->thr)) BAD("3");
      if (!seg->vts) BAD("4");
      if (seg->prev && seg->prev->vts
          && !cmpGEQ_VTS(seg->vts, seg->prev->vts))
         BAD("5");
      if (seg->other && seg->other->vts
          && !cmpGEQ_VTS(seg->vts, seg->other->vts))
         BAD("6");
   }
   return;
  bad:
   VG_(printf)("segments__sanity_check: who=\"%s\", bad=\"%s\"\n", 
               who, how);
   tl_assert(0);
#undef BAD
}


/* Sanity check shadow memory, as far as possible */
static Int cmp_Addr_for_ssort ( void* p1, void* p2 ) {
   Addr a1 = *(Addr*)p1;
   Addr a2 = *(Addr*)p2;
   if (a1 < a2) return -1;
   if (a1 > a2) return 1;
   return 0;
}
__attribute__((noinline))
static void shmem__sanity_check ( Char* who )
{
#define BAD(_str) do { how = (_str); goto bad; } while (0)
   Char*   how = "no error";
   Word    smga;
   SecMap* sm;
   Word    i, j, ws_size, n_valid_tags;
   Word*   ws_words;
   Addr*   valid_tags;
   HG_(initIterFM)( map_shmem );
   // for sm in SecMaps {
   while (HG_(nextIterFM)( map_shmem,
                           (Word*)(void*)&smga, (Word*)(void*)&sm )) {
      SecMapIter itr;
      UInt*      w32p;
      Bool       mbHasShared = False;
      Bool       allNoAccess = True;
      if (!is_sane_SecMap(sm)) BAD("1");
      // sm properly aligned
      if (smga != shmem__round_to_SecMap_base(smga)) BAD("2");
      // if any shadow word is ShR or ShM then .mbHasShared == True
      initSecMapIter( &itr );
      while (stepSecMapIter( &w32p, &itr, sm )) {
         UInt w32 = *w32p;
         if (is_SHVAL_Sh(w32)) 
            mbHasShared = True;
         if (!is_SHVAL_NoAccess(w32))
            allNoAccess = False;
         if (is_SHVAL_Excl(w32)) {
            // for each Excl(segid) state
            // map_segments_lookup maps to a sane Segment(seg)
            Segment*  seg;
            SegmentID segid = un_SHVAL_Excl(w32);
            if (!is_sane_SegmentID(segid)) BAD("3");
            seg = map_segments_maybe_lookup(segid);
            if (!is_sane_Segment(seg)) BAD("4");
         } 
         else if (is_SHVAL_Sh(w32)) {
            WordSetID tset = un_SHVAL_Sh_tset(w32);
            WordSetID lset = un_SHVAL_Sh_lset(w32);
            if (!HG_(plausibleWS)( univ_tsets, tset )) BAD("5");
            if (!HG_(saneWS_SLOW)( univ_tsets, tset )) BAD("6");
            if (HG_(cardinalityWS)( univ_tsets, tset ) < 2) BAD("7");
            if (!HG_(plausibleWS)( univ_lsets, lset )) BAD("8");
            if (!HG_(saneWS_SLOW)( univ_lsets, lset )) BAD("9");
            HG_(getPayloadWS)( &ws_words, &ws_size, univ_lsets, lset );
            for (j = 0; j < ws_size; j++) {
               Lock* lk = (Lock*)ws_words[j];
               // for each ShM/ShR(tsetid,lsetid) state
               // each lk in lset is a valid Lock
               if (!is_sane_LockN(lk)) BAD("10");
            }
            HG_(getPayloadWS)( &ws_words, &ws_size, univ_tsets, tset );
            for (j = 0; j < ws_size; j++) {
               Thread* thr = (Thread*)ws_words[j];
               //for each ShM/ShR(tsetid,lsetid) state
               // each thr in tset is a valid thread, which is non-dead
               if (!is_sane_Thread(thr)) BAD("11");
            }
         }
         else if (is_SHVAL_NoAccess(w32) || is_SHVAL_New(w32)) {
            /* nothing to check */
         }
         else {
            /* bogus shadow mem value */
            BAD("12");
         }
      } /* iterating over a SecMap */
      // Check essential safety property
      if (mbHasShared && !sm->mbHasShared) BAD("13");
      // This is optional - check that destroyed memory has its hint
      // bits cleared.  NB won't work properly unless full, eager
      // GCing of SecMaps is implemented
      //if (allNoAccess && sm->mbHasLocks) BAD("13a");
   }
   HG_(doneIterFM)( map_shmem );

   // check the cache
   valid_tags   = hg_zalloc(N_WAY_NENT * sizeof(Addr));
   n_valid_tags = 0;
   tl_assert(valid_tags);
   for (i = 0; i < N_WAY_NENT; i++) {
      CacheLine* cl;
      Addr       tag; 
      /* way0, dude */
      cl  = &cache_shmem.lyns0[i];
      tag =  cache_shmem.tags0[i];
      if (tag != 1) {
         if (!is_valid_scache_tag(tag)) BAD("14-0");
         if (!is_sane_CacheLine(cl)) BAD("15-0");
         /* A valid tag should be of the form 
            X---X line_number:N_WAY_BITS 0:N_LINE_BITS */
         if (tag & (N_LINE_ARANGE-1)) BAD("16-0");
         if ( i != ((tag >> N_LINE_BITS) & (N_WAY_NENT-1)) ) BAD("16-1");
         valid_tags[n_valid_tags++] = tag;
      }
   }
   tl_assert(n_valid_tags <= N_WAY_NENT);
   if (n_valid_tags > 1) {
      /* Check that the valid tags are unique */
      VG_(ssort)( valid_tags, n_valid_tags, sizeof(Addr), cmp_Addr_for_ssort );
      for (i = 0; i < n_valid_tags-1; i++) {
         if (valid_tags[i] >= valid_tags[i+1])
            BAD("16-2");
      }
   }
   hg_free(valid_tags);
   return;
  bad:
   VG_(printf)("shmem__sanity_check: who=\"%s\", bad=\"%s\"\n", who, how);
   tl_assert(0);
#undef BAD
}

static void all_except_Locks__sanity_check ( Char* who ) {
   stats__sanity_checks++;
   if (0) VG_(printf)("all_except_Locks__sanity_check(%s)\n", who);
   threads__sanity_check(who);
   segments__sanity_check(who);
   shmem__sanity_check(who);
   laog__sanity_check(who);
}
static void all__sanity_check ( Char* who ) {
   all_except_Locks__sanity_check(who);
   locks__sanity_check(who);
}


/*----------------------------------------------------------------*/
/*--- the core memory state machine (msm__* functions)         ---*/
/*----------------------------------------------------------------*/

static UWord stats__msm_read_Excl_nochange = 0;
static UWord stats__msm_read_Excl_transfer = 0;
static UWord stats__msm_read_Excl_to_ShR   = 0;
static UWord stats__msm_read_ShR_to_ShR    = 0;
static UWord stats__msm_read_ShM_to_ShM    = 0;
static UWord stats__msm_read_New_to_Excl   = 0;
static UWord stats__msm_read_NoAccess      = 0;

static UWord stats__msm_write_Excl_nochange = 0;
static UWord stats__msm_write_Excl_transfer = 0;
static UWord stats__msm_write_Excl_to_ShM   = 0;
static UWord stats__msm_write_ShR_to_ShM    = 0;
static UWord stats__msm_write_ShM_to_ShM    = 0;
static UWord stats__msm_write_New_to_Excl   = 0;
static UWord stats__msm_write_NoAccess      = 0;

/* fwds */
static void record_error_Race ( Thread* thr, 
                                Addr data_addr, Bool isWrite, Int szB,
                                UInt old_sv, UInt new_sv, 
                                ExeContext* mb_lastlock );

static void record_error_FreeMemLock ( Thread* thr, Lock* lk );

static void record_error_UnlockUnlocked ( Thread*, Lock* );
static void record_error_UnlockForeign  ( Thread*, Thread*, Lock* );
static void record_error_UnlockBogus    ( Thread*, Addr );
static void record_error_PthAPIerror    ( Thread*, HChar*, Word, HChar* );
static void record_error_LockOrder      ( Thread*, Addr, Addr,
                                                   ExeContext*, ExeContext* );

static void record_error_Misc ( Thread*, HChar* );
static void announce_one_thread ( Thread* thr ); /* fwds */

static WordSetID add_BHL ( WordSetID lockset ) {
   return HG_(addToWS)( univ_lsets, lockset, (Word)__bus_lock_Lock );
}
static WordSetID del_BHL ( WordSetID lockset ) {
   return HG_(delFromWS)( univ_lsets, lockset, (Word)__bus_lock_Lock );
}


/* Last-lock-lossage records.  This mechanism exists to help explain
   to programmers why we are complaining about a race.  The idea is to
   monitor all lockset transitions.  When a previously nonempty
   lockset becomes empty, the lock(s) that just disappeared (the
   "lossage") are the locks that have consistently protected the
   location (ga_of_access) in question for the longest time.  Most of
   the time the lossage-set is a single lock.  Because the
   lossage-lock is the one that has survived longest, there is there
   is a good chance that it is indeed the lock that the programmer
   intended to use to protect the location.

   Note that we cannot in general just look at the lossage set when we
   see a transition to ShM(...,empty-set), because a transition to an
   empty lockset can happen arbitrarily far before the point where we
   want to report an error.  This is in the case where there are many
   transitions ShR -> ShR, all with an empty lockset, and only later
   is there a transition to ShM.  So what we want to do is note the
   lossage lock at the point where a ShR -> ShR transition empties out
   the lockset, so we can present it later if there should be a
   transition to ShM.

   So this function finds such transitions.  For each, it associates
   in ga_to_lastlock, the guest address and the lossage lock.  In fact
   we do not record the Lock* directly as that may disappear later,
   but instead the ExeContext inside the Lock which says where it was
   initialised or first locked.  ExeContexts are permanent so keeping
   them indefinitely is safe.

   A boring detail: the hardware bus lock is not interesting in this
   respect, so we first remove that from the pre/post locksets.
*/

static UWord stats__ga_LL_adds = 0;

static WordFM* ga_to_lastlock = NULL; /* GuestAddr -> ExeContext* */

static 
void record_last_lock_lossage ( Addr ga_of_access,
                                WordSetID lset_old, WordSetID lset_new )
{
   Lock* lk;
   Int   card_old, card_new;

   tl_assert(lset_old != lset_new);

   if (0) VG_(printf)("XX1: %d (card %d) -> %d (card %d) %p\n", 
                      (Int)lset_old, 
                      HG_(cardinalityWS)(univ_lsets,lset_old),
                      (Int)lset_new, 
                      HG_(cardinalityWS)(univ_lsets,lset_new),
                      ga_of_access );

   /* This is slow, but at least it's simple.  The bus hardware lock
      just confuses the logic, so remove it from the locksets we're
      considering before doing anything else. */
   lset_new = del_BHL( lset_new );

   if (!HG_(isEmptyWS)( univ_lsets, lset_new )) {
      /* The post-transition lock set is not empty.  So we are not
         interested.  We're only interested in spotting transitions
         that make locksets become empty. */
      return;
   }

   /* lset_new is now empty */
   card_new = HG_(cardinalityWS)( univ_lsets, lset_new );
   tl_assert(card_new == 0);

   lset_old = del_BHL( lset_old );
   card_old = HG_(cardinalityWS)( univ_lsets, lset_old );

   if (0) VG_(printf)(" X2: %d (card %d) -> %d (card %d)\n",
                      (Int)lset_old, card_old, (Int)lset_new, card_new );

   if (card_old == 0) {
      /* The old lockset was also empty.  Not interesting. */
      return;
   }

   tl_assert(card_old > 0);
   tl_assert(!HG_(isEmptyWS)( univ_lsets, lset_old ));

   /* Now we know we've got a transition from a nonempty lockset to an
      empty one.  So lset_old must be the set of locks lost.  Record
      some details.  If there is more than one element in the lossage
      set, just choose one arbitrarily -- not the best, but at least
      it's simple. */

   lk = (Lock*)HG_(anyElementOfWS)( univ_lsets, lset_old );
   if (0) VG_(printf)("lossage %d %p\n", 
                      HG_(cardinalityWS)( univ_lsets, lset_old), lk );
   if (lk->appeared_at) {
      if (ga_to_lastlock == NULL)
         ga_to_lastlock = HG_(newFM)( hg_zalloc, hg_free, NULL );
      HG_(addToFM)( ga_to_lastlock, ga_of_access, (Word)lk->appeared_at );
      stats__ga_LL_adds++;
   }
}

/* This queries the table (ga_to_lastlock) made by
   record_last_lock_lossage, when constructing error messages.  It
   attempts to find the ExeContext of the allocation or initialisation
   point for the lossage lock associated with 'ga'. */

static ExeContext* maybe_get_lastlock_initpoint ( Addr ga ) 
{
   ExeContext* ec_hint = NULL;
   if (ga_to_lastlock != NULL 
       && HG_(lookupFM)(ga_to_lastlock, 
                        NULL, (Word*)(void*)&ec_hint, ga)) {
      tl_assert(ec_hint != NULL);
      return ec_hint;
   } else {
      return NULL;
   }
}


static void msm__show_state_change ( Thread* thr_acc, Addr a, Int szB,
                                     Char howC,
                                     UInt sv_old, UInt sv_new )
{
   ThreadId tid;
   UChar txt_old[100], txt_new[100];
   Char* how = "";
   tl_assert(is_sane_Thread(thr_acc));
   tl_assert(clo_trace_level == 1 || clo_trace_level == 2);
   switch (howC) {
      case 'r': how = "rd"; break;
      case 'w': how = "wr"; break;
      case 'p': how = "pa"; break;
      default: tl_assert(0);
   }
   show_shadow_w32_for_user(txt_old, sizeof(txt_old), sv_old);
   show_shadow_w32_for_user(txt_new, sizeof(txt_new), sv_new);
   txt_old[sizeof(txt_old)-1] = 0;
   txt_new[sizeof(txt_new)-1] = 0;
   if (clo_trace_level == 2) {
      /* show everything */
      VG_(message)(Vg_UserMsg, "");
      announce_one_thread( thr_acc );
      VG_(message)(Vg_UserMsg, 
                   "TRACE: %p %s %d thr#%d :: %s --> %s",
                   a, how, szB, thr_acc->errmsg_index, txt_old, txt_new );
      tid = map_threads_maybe_reverse_lookup_SLOW(thr_acc);
      if (tid != VG_INVALID_THREADID) {
         VG_(get_and_pp_StackTrace)( tid, 8 );
      }
   } else {
      /* Just print one line */
      VG_(message)(Vg_UserMsg, 
                   "TRACE: %p %s %d thr#%d :: %22s --> %22s",
                   a, how, szB, thr_acc->errmsg_index, txt_old, txt_new );
   }
}


/* Here are some MSM stats from startup/shutdown of OpenOffice.

     msm:  489,734,723   80,278,862 rd/wr_Excl_nochange
     msm:    3,171,542       93,738 rd/wr_Excl_transfer
     msm:       45,036          167 rd/wr_Excl_to_ShR/ShM
     msm:   13,352,594          285 rd/wr_ShR_to_ShR/ShM
     msm:    1,125,879      815,779 rd/wr_ShM_to_ShM
     msm:    7,561,842  250,629,935 rd/wr_New_to_Excl
     msm:       17,778            0 rd/wr_NoAccess

   This says how the clauses should be ordered for greatest speed:

   * the vast majority of memory reads (490 million out of a total of
     515 million) are of memory in an exclusive state, and the state
     is unchanged.  All other read accesses are insignificant by
     comparison.

   * 75% (251 million out of a total of 332 million) writes are 'first
     time' writes, which take New memory into exclusive ownership.
     Almost all the rest (80 million) are accesses to exclusive state,
     which remains unchanged.  All other write accesses are
     insignificant. */

/* The core MSM.  If 'wold' is the old 32-bit shadow word for a
   location, return the new shadow word that would result for a read
   of the location, and report any errors necessary on the way.  This
   does not update shadow memory - it merely produces new shadow words
   from old.  'thr_acc' and 'a' are supplied only so it can produce
   coherent error messages if necessary. */
static
UInt msm__handle_read ( Thread* thr_acc, Addr a, UInt wold, Int szB )
{
   UInt wnew = SHVAL_Invalid;

   tl_assert(is_sane_Thread(thr_acc));

   if (0) VG_(printf)("read thr=%p %p\n", thr_acc, a);

   /* Exclusive */
   if (LIKELY(is_SHVAL_Excl(wold))) {
      /* read Excl(segid) 
           |  segid_old == segid-of-thread
           -> no change
           |  segid_old `happens_before` segid-of-this-thread
           -> Excl(segid-of-this-thread)
           |  otherwise
           -> ShR
      */
      SegmentID segid_old = un_SHVAL_Excl(wold);
      tl_assert(is_sane_SegmentID(segid_old));
      if (LIKELY(segid_old == thr_acc->csegid)) {
         /* no change */
         stats__msm_read_Excl_nochange++;
         /*NOCHANGE*/return wold;
      }
      if (happens_before(segid_old, thr_acc->csegid)) {
         /* -> Excl(segid-of-this-thread) */
         wnew = mk_SHVAL_Excl(thr_acc->csegid);
         stats__msm_read_Excl_transfer++;
         goto changed;
      }
      /* else */ {
         /* Enter the shared-readonly (ShR) state. */
         WordSetID tset, lset;
         /* This location has been accessed by precisely two threads.
            Make an appropriate tset. */
         // FIXME: performance: duplicate map_segments_lookup(segid_old)
         // since must also be done in happens_before()
         Segment* seg_old = map_segments_lookup( segid_old );
         Thread*  thr_old = seg_old->thr;
         tset = HG_(doubletonWS)( univ_tsets, (Word)thr_old, (Word)thr_acc );
         lset = add_BHL( thr_acc->locksetA ); /* read ==> use all locks */
         wnew = mk_SHVAL_ShR( tset, lset );
         stats__msm_read_Excl_to_ShR++;
         goto changed;
      }
      /*NOTREACHED*/
   } 

   /* Shared-Readonly */
   if (is_SHVAL_ShR(wold)) {
     /* read Shared-Readonly(threadset, lockset)
        We remain in ShR state, but add this thread to the 
        threadset and refine the lockset accordingly.  Do not
        complain if the lockset becomes empty -- that's ok. */
      WordSetID tset_old = un_SHVAL_ShR_tset(wold);
      WordSetID lset_old = un_SHVAL_ShR_lset(wold);
      WordSetID tset_new = HG_(addToWS)( univ_tsets, 
                                         tset_old, (Word)thr_acc );
      WordSetID lset_new = HG_(intersectWS)( univ_lsets,
                                             lset_old, 
                                             add_BHL(thr_acc->locksetA)
                                             /* read ==> use all locks */ );
      /*UInt*/  wnew     = mk_SHVAL_ShR( tset_new, lset_new );
      if (lset_old != lset_new)
         record_last_lock_lossage(a,lset_old,lset_new);
      stats__msm_read_ShR_to_ShR++;
      goto changed;
   }

   /* Shared-Modified */
   if (is_SHVAL_ShM(wold)) {
      /* read Shared-Modified(threadset, lockset)
         We remain in ShM state, but add this thread to the 
         threadset and refine the lockset accordingly.
         If the lockset becomes empty, complain. */
      WordSetID tset_old = un_SHVAL_ShM_tset(wold);
      WordSetID lset_old = un_SHVAL_ShM_lset(wold);
      WordSetID tset_new = HG_(addToWS)( univ_tsets,
                                         tset_old, (Word)thr_acc );
      WordSetID lset_new = HG_(intersectWS)( univ_lsets,
                                             lset_old,
                                             add_BHL(thr_acc->locksetA)
                                             /* read ==> use all locks */ ); 
      /*UInt*/  wnew     = mk_SHVAL_ShM( tset_new, lset_new );
      if (lset_old != lset_new)
         record_last_lock_lossage(a,lset_old,lset_new);
      if (HG_(isEmptyWS)(univ_lsets, lset_new)
          && !HG_(isEmptyWS)(univ_lsets, lset_old)) {
         record_error_Race( thr_acc, a, 
                            False/*isWrite*/, szB, wold, wnew,
                            maybe_get_lastlock_initpoint(a) );
      }
      stats__msm_read_ShM_to_ShM++;
      goto changed;
   }
 
   /* New */
   if (is_SHVAL_New(wold)) {
      /* read New -> Excl(segid) */
      wnew = mk_SHVAL_Excl( thr_acc->csegid );
      stats__msm_read_New_to_Excl++;
      goto changed;
   } 

   /* NoAccess */
   if (is_SHVAL_NoAccess(wold)) {
      // FIXME: complain if accessing here
      // FIXME: transition to Excl?
      if (0)
      VG_(printf)(
         "msm__handle_read_aligned_32(thr=%p, addr=%p): NoAccess\n",
         thr_acc, (void*)a );
      stats__msm_read_NoAccess++;
      /*NOCHANGE*/return wold; /* no change */
   }

   /* hmm, bogus state */
   tl_assert(0);

  changed:
   if (UNLIKELY(clo_trace_level > 0)) {
      if (a <= clo_trace_addr && clo_trace_addr < a+szB
          && wold != wnew) {
         msm__show_state_change( thr_acc, a, szB, 'r', wold, wnew );
      }
   }
   return wnew;
}

/* Similar to msm__handle_read, compute a new 32-bit shadow word
   resulting from a write to a location, and report any errors
   necessary on the way. */
static
UInt msm__handle_write ( Thread* thr_acc, Addr a, UInt wold, Int szB )
{
   UInt wnew = SHVAL_Invalid;

   tl_assert(is_sane_Thread(thr_acc));

   if (0) VG_(printf)("write32 thr=%p %p\n", thr_acc, a);

   /* New */
   if (LIKELY(is_SHVAL_New(wold))) {
      /* write New -> Excl(segid) */
      wnew = mk_SHVAL_Excl( thr_acc->csegid );
      stats__msm_write_New_to_Excl++;
      goto changed;
   }

   /* Exclusive */
   if (is_SHVAL_Excl(wold)) {
      // I believe is identical to case for read Excl
      // apart from enters ShM rather than ShR 
      /* read Excl(segid) 
           |  segid_old == segid-of-thread
           -> no change
           |  segid_old `happens_before` segid-of-this-thread
           -> Excl(segid-of-this-thread)
           |  otherwise
           -> ShM
      */
      SegmentID segid_old = un_SHVAL_Excl(wold);
      tl_assert(is_sane_SegmentID(segid_old));
      if (segid_old == thr_acc->csegid) {
         /* no change */
         stats__msm_write_Excl_nochange++;
         /*NOCHANGE*/return wold;
      }
      if (happens_before(segid_old, thr_acc->csegid)) {
         /* -> Excl(segid-of-this-thread) */
         wnew = mk_SHVAL_Excl(thr_acc->csegid);
         stats__msm_write_Excl_transfer++;
         goto changed;
      }
      /* else */ {
         /* Enter the shared-modified (ShM) state. */
         WordSetID tset, lset;
         /* This location has been accessed by precisely two threads.
            Make an appropriate tset. */
         // FIXME: performance: duplicate map_segments_lookup(segid_old)
         // since must also be done in happens_before()
         Segment* seg_old = map_segments_lookup( segid_old );
         Thread*  thr_old = seg_old->thr;
         tset = HG_(doubletonWS)( univ_tsets, (Word)thr_old, (Word)thr_acc );
         lset = thr_acc->locksetW; /* write ==> use only w-held locks */
         wnew = mk_SHVAL_ShM( tset, lset );
         if (HG_(isEmptyWS)(univ_lsets, lset)) {
            record_error_Race( thr_acc, 
                               a, True/*isWrite*/, szB, wold, wnew,
                               maybe_get_lastlock_initpoint(a) );
         }
         stats__msm_write_Excl_to_ShM++;
         goto changed;
      }
      /*NOTREACHED*/
   } 

   /* Shared-Readonly */
   if (is_SHVAL_ShR(wold)) {
      /* write Shared-Readonly(threadset, lockset)
         We move to ShM state, add this thread to the 
         threadset and refine the lockset accordingly.
         If the lockset becomes empty, complain. */
      WordSetID tset_old = un_SHVAL_ShR_tset(wold);
      WordSetID lset_old = un_SHVAL_ShR_lset(wold);
      WordSetID tset_new = HG_(addToWS)( univ_tsets, 
                                         tset_old, (Word)thr_acc );
      WordSetID lset_new = HG_(intersectWS)(
                              univ_lsets, 
                              lset_old, 
                              thr_acc->locksetW
                              /* write ==> use only w-held locks */
                           );
      /*UInt*/  wnew     = mk_SHVAL_ShM( tset_new, lset_new );
      if (lset_old != lset_new)
         record_last_lock_lossage(a,lset_old,lset_new);
      if (HG_(isEmptyWS)(univ_lsets, lset_new)) {
         record_error_Race( thr_acc, a, 
                            True/*isWrite*/, szB, wold, wnew,
                            maybe_get_lastlock_initpoint(a) );
      }
      stats__msm_write_ShR_to_ShM++;
      goto changed;
   }

   /* Shared-Modified */
   else if (is_SHVAL_ShM(wold)) {
      /* write Shared-Modified(threadset, lockset)
         We remain in ShM state, but add this thread to the 
         threadset and refine the lockset accordingly.
         If the lockset becomes empty, complain. */
      WordSetID tset_old = un_SHVAL_ShM_tset(wold);
      WordSetID lset_old = un_SHVAL_ShM_lset(wold);
      WordSetID tset_new = HG_(addToWS)( univ_tsets,
                                         tset_old, (Word)thr_acc );
      WordSetID lset_new = HG_(intersectWS)( 
                              univ_lsets,
                              lset_old, 
                              thr_acc->locksetW 
                              /* write ==> use only w-held locks */
                           ); 
      /*UInt*/  wnew     = mk_SHVAL_ShM( tset_new, lset_new );
      if (lset_old != lset_new)
         record_last_lock_lossage(a,lset_old,lset_new);
      if (HG_(isEmptyWS)(univ_lsets, lset_new)
          && !HG_(isEmptyWS)(univ_lsets, lset_old)) {
         record_error_Race( thr_acc, a, 
                            True/*isWrite*/, szB, wold, wnew,
                            maybe_get_lastlock_initpoint(a) );
      }
      stats__msm_write_ShM_to_ShM++;
      goto changed;
   }

   /* NoAccess */
   if (is_SHVAL_NoAccess(wold)) {
      // FIXME: complain if accessing here
      // FIXME: transition to Excl?
      if (0)
      VG_(printf)(
         "msm__handle_write_aligned_32(thr=%p, addr=%p): NoAccess\n",
         thr_acc, (void*)a );
      stats__msm_write_NoAccess++;
      /*NOCHANGE*/return wold;
   } 

   /* hmm, bogus state */
   VG_(printf)("msm__handle_write_aligned_32: bogus old state 0x%x\n", 
               wold);
   tl_assert(0);

  changed:
   if (UNLIKELY(clo_trace_level > 0)) {
      if (a <= clo_trace_addr && clo_trace_addr < a+szB
          && wold != wnew) {
         msm__show_state_change( thr_acc, a, szB, 'w', wold, wnew );
      }
   }
   return wnew;
}


/*----------------------------------------------------------------*/
/*--- Shadow value and address range handlers                  ---*/
/*----------------------------------------------------------------*/

static void laog__pre_thread_acquires_lock ( Thread*, Lock* ); /* fwds */
static void laog__handle_lock_deletions    ( WordSetID ); /* fwds */
static inline Thread* get_current_Thread ( void ); /* fwds */

/* ------------ CacheLineF and CacheLineZ related ------------ */

static void write_twobit_array ( UChar* arr, UWord ix, UWord b2 ) {
   Word bix, shft, mask, prep;
   tl_assert((b2 & ~3) == 0);
   tl_assert(ix >= 0);
   bix  = ix >> 2;
   shft = 2 * (ix & 3); /* 0, 2, 4 or 6 */
   mask = 3 << shft;
   prep = b2 << shft;
   arr[bix] = (arr[bix] & ~mask) | prep;
}

static UWord read_twobit_array ( UChar* arr, UWord ix ) {
   Word bix, shft;
   tl_assert(ix >= 0);
   bix  = ix >> 2;
   shft = 2 * (ix & 3); /* 0, 2, 4 or 6 */
   return (arr[bix] >> shft) & 3;
}

/* Given a lineZ index and a SecMap, return the CacheLineZ* and CacheLineF*
   for that index. */
static void get_ZF_by_index ( /*OUT*/CacheLineZ** zp,
                              /*OUT*/CacheLineF** fp,
                              SecMap* sm, Int zix ) {
   CacheLineZ* lineZ;
   tl_assert(zp);
   tl_assert(fp);
   tl_assert(zix >= 0 && zix < N_SECMAP_ZLINES);
   tl_assert(is_sane_SecMap(sm));
   lineZ = &sm->linesZ[zix];
   if (lineZ->dict[0] == 0) {
      Int fix = lineZ->dict[1];
      tl_assert(sm->linesF);
      tl_assert(sm->linesF_size > 0);
      tl_assert(fix >= 0 && fix < sm->linesF_size);
      *zp = NULL;
      *fp = &sm->linesF[fix];
      tl_assert(sm->linesF[fix].inUse);
   } else {
      *zp = lineZ;
      *fp = NULL;
   }
}

static void find_ZF_for_reading ( /*OUT*/CacheLineZ** zp,
                                  /*OUT*/CacheLineF** fp, Addr tag ) {
   CacheLineZ* lineZ;
   CacheLineF* lineF;
   UWord   zix;
   SecMap* sm    = shmem__find_or_alloc_SecMap(tag);
   UWord   smoff = shmem__get_SecMap_offset(tag);
   /* since smoff is derived from a valid tag, it should be
      cacheline-aligned. */
   tl_assert(0 == (smoff & (N_LINE_ARANGE - 1)));
   zix = smoff >> N_LINE_BITS;
   tl_assert(zix < N_SECMAP_ZLINES);
   lineZ = &sm->linesZ[zix];
   lineF = NULL;
   if (lineZ->dict[0] == 0) {
      Word fix = lineZ->dict[1];
      tl_assert(sm->linesF);
      tl_assert(sm->linesF_size > 0);
      tl_assert(fix >= 0 && fix < sm->linesF_size);
      lineF = &sm->linesF[fix];
      tl_assert(lineF->inUse);
      lineZ = NULL;
   }
   *zp = lineZ;
   *fp = lineF;
}

static void find_Z_for_writing ( /*OUT*/SecMap** smp,
                                 /*OUT*/Word* zixp,
                                 Addr tag ) {
   CacheLineZ* lineZ;
   CacheLineF* lineF;
   UWord   zix;
   SecMap* sm    = shmem__find_or_alloc_SecMap(tag);
   UWord   smoff = shmem__get_SecMap_offset(tag);
   /* since smoff is derived from a valid tag, it should be
      cacheline-aligned. */
   tl_assert(0 == (smoff & (N_LINE_ARANGE - 1)));
   zix = smoff >> N_LINE_BITS;
   tl_assert(zix < N_SECMAP_ZLINES);
   lineZ = &sm->linesZ[zix];
   lineF = NULL;
   /* If lineZ has an associated lineF, free it up. */
   if (lineZ->dict[0] == 0) {
      Word fix = lineZ->dict[1];
      tl_assert(sm->linesF);
      tl_assert(sm->linesF_size > 0);
      tl_assert(fix >= 0 && fix < sm->linesF_size);
      lineF = &sm->linesF[fix];
      tl_assert(lineF->inUse);
      lineF->inUse = False;
   }
   *smp  = sm;
   *zixp = zix;
}

static 
void alloc_F_for_writing ( /*MOD*/SecMap* sm, /*OUT*/Word* fixp ) {
   Word        i, new_size;
   CacheLineF* nyu;

   if (sm->linesF) {
      tl_assert(sm->linesF_size > 0);
   } else {
      tl_assert(sm->linesF_size == 0);
   }

   if (sm->linesF) {
      for (i = 0; i < sm->linesF_size; i++) {
         if (!sm->linesF[i].inUse) {
            *fixp = (Word)i;
            return;
         }
      }
   }

   /* No free F line found.  Expand existing array and try again. */
   new_size = sm->linesF_size==0 ? 1 : 2 * sm->linesF_size;
   nyu      = hg_zalloc( new_size * sizeof(CacheLineF) );
   tl_assert(nyu);

   stats__secmap_linesF_allocd += (new_size - sm->linesF_size);
   stats__secmap_linesF_bytes  += (new_size - sm->linesF_size)
                                  * sizeof(CacheLineF);

   if (0)
   VG_(printf)("SM %p: expand F array from %d to %d\n", 
               sm, (Int)sm->linesF_size, new_size);

   for (i = 0; i < new_size; i++)
      nyu[i].inUse = False;

   if (sm->linesF) {
      for (i = 0; i < sm->linesF_size; i++) {
         tl_assert(sm->linesF[i].inUse);
         nyu[i] = sm->linesF[i];
      }
      VG_(memset)(sm->linesF, 0, sm->linesF_size * sizeof(CacheLineF) );
      hg_free(sm->linesF);
   }

   sm->linesF      = nyu;
   sm->linesF_size = new_size;

   for (i = 0; i < sm->linesF_size; i++) {
      if (!sm->linesF[i].inUse) {
         *fixp = (Word)i;
         return;
      }
    }

    /*NOTREACHED*/
    tl_assert(0);
}


/* ------------ CacheLine and implicit-tree related ------------ */

__attribute__((unused))
static void pp_CacheLine ( CacheLine* cl ) {
   Word i;
   if (!cl) {
      VG_(printf)("pp_CacheLine(NULL)\n");
      return;
   }
   for (i = 0; i < N_LINE_TREES; i++) 
      VG_(printf)("   descr: %04lx\n", (UWord)cl->descrs[i]);
   for (i = 0; i < N_LINE_ARANGE; i++) 
      VG_(printf)("    sval: %08lx\n", (UWord)cl->svals[i]);
}

static UChar descr_to_validbits ( UShort descr )
{
   /* a.k.a Party Time for gcc's constant folder */
#  define DESCR(b8_7, b8_6, b8_5, b8_4, b8_3, b8_2, b8_1, b8_0, \
                b16_3, b32_1, b16_2, b64, b16_1, b32_0, b16_0)  \
             ( (UShort) ( ( (b8_7)  << 14) | ( (b8_6)  << 13) | \
                          ( (b8_5)  << 12) | ( (b8_4)  << 11) | \
                          ( (b8_3)  << 10) | ( (b8_2)  << 9)  | \
                          ( (b8_1)  << 8)  | ( (b8_0)  << 7)  | \
                          ( (b16_3) << 6)  | ( (b32_1) << 5)  | \
                          ( (b16_2) << 4)  | ( (b64)   << 3)  | \
                          ( (b16_1) << 2)  | ( (b32_0) << 1)  | \
                          ( (b16_0) << 0) ) )

#  define BYTE(bit7, bit6, bit5, bit4, bit3, bit2, bit1, bit0) \
             ( (UChar) ( ( (bit7) << 7) | ( (bit6) << 6) | \
                         ( (bit5) << 5) | ( (bit4) << 4) | \
                         ( (bit3) << 3) | ( (bit2) << 2) | \
                         ( (bit1) << 1) | ( (bit0) << 0) ) )

   /* these should all get folded out at compile time */
   tl_assert(DESCR(1,0,0,0,0,0,0,0, 0,0,0, 0, 0,0,0) == TREE_DESCR_8_7);
   tl_assert(DESCR(0,0,0,0,0,0,0,1, 0,0,0, 0, 0,0,0) == TREE_DESCR_8_0);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 1,0,0, 0, 0,0,0) == TREE_DESCR_16_3);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,1,0, 0, 0,0,0) == TREE_DESCR_32_1);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,0,1, 0, 0,0,0) == TREE_DESCR_16_2);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,0,0, 1, 0,0,0) == TREE_DESCR_64);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,0,0, 0, 1,0,0) == TREE_DESCR_16_1);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,0,0, 0, 0,1,0) == TREE_DESCR_32_0);
   tl_assert(DESCR(0,0,0,0,0,0,0,0, 0,0,0, 0, 0,0,1) == TREE_DESCR_16_0);

   switch (descr) {
   /*
              +--------------------------------- TREE_DESCR_8_7
              |             +------------------- TREE_DESCR_8_0
              |             |  +---------------- TREE_DESCR_16_3
              |             |  | +-------------- TREE_DESCR_32_1
              |             |  | | +------------ TREE_DESCR_16_2
              |             |  | | |  +--------- TREE_DESCR_64
              |             |  | | |  |  +------ TREE_DESCR_16_1
              |             |  | | |  |  | +---- TREE_DESCR_32_0
              |             |  | | |  |  | | +-- TREE_DESCR_16_0
              |             |  | | |  |  | | |
              |             |  | | |  |  | | |   GRANULARITY, 7 -> 0 */
   case DESCR(1,1,1,1,1,1,1,1, 0,0,0, 0, 0,0,0): /* 8 8 8 8  8 8 8 8 */
                                                 return BYTE(1,1,1,1,1,1,1,1);
   case DESCR(1,1,0,0,1,1,1,1, 0,0,1, 0, 0,0,0): /* 8 8 16   8 8 8 8 */
                                                 return BYTE(1,1,0,1,1,1,1,1);
   case DESCR(0,0,1,1,1,1,1,1, 1,0,0, 0, 0,0,0): /* 16  8 8  8 8 8 8 */ 
                                                 return BYTE(0,1,1,1,1,1,1,1);
   case DESCR(0,0,0,0,1,1,1,1, 1,0,1, 0, 0,0,0): /* 16  16   8 8 8 8 */
                                                 return BYTE(0,1,0,1,1,1,1,1);

   case DESCR(1,1,1,1,1,1,0,0, 0,0,0, 0, 0,0,1): /* 8 8 8 8  8 8 16 */ 
                                                 return BYTE(1,1,1,1,1,1,0,1);
   case DESCR(1,1,0,0,1,1,0,0, 0,0,1, 0, 0,0,1): /* 8 8 16   8 8 16 */
                                                 return BYTE(1,1,0,1,1,1,0,1);
   case DESCR(0,0,1,1,1,1,0,0, 1,0,0, 0, 0,0,1): /* 16  8 8  8 8 16 */
                                                 return BYTE(0,1,1,1,1,1,0,1);
   case DESCR(0,0,0,0,1,1,0,0, 1,0,1, 0, 0,0,1): /* 16  16   8 8 16 */
                                                 return BYTE(0,1,0,1,1,1,0,1);

   case DESCR(1,1,1,1,0,0,1,1, 0,0,0, 0, 1,0,0): /* 8 8 8 8  16 8 8 */
                                                 return BYTE(1,1,1,1,0,1,1,1);
   case DESCR(1,1,0,0,0,0,1,1, 0,0,1, 0, 1,0,0): /* 8 8 16   16 8 8 */
                                                 return BYTE(1,1,0,1,0,1,1,1);
   case DESCR(0,0,1,1,0,0,1,1, 1,0,0, 0, 1,0,0): /* 16  8 8  16 8 8 */
                                                 return BYTE(0,1,1,1,0,1,1,1);
   case DESCR(0,0,0,0,0,0,1,1, 1,0,1, 0, 1,0,0): /* 16  16   16 8 8 */
                                                 return BYTE(0,1,0,1,0,1,1,1);

   case DESCR(1,1,1,1,0,0,0,0, 0,0,0, 0, 1,0,1): /* 8 8 8 8  16 16 */
                                                 return BYTE(1,1,1,1,0,1,0,1);
   case DESCR(1,1,0,0,0,0,0,0, 0,0,1, 0, 1,0,1): /* 8 8 16   16 16 */
                                                 return BYTE(1,1,0,1,0,1,0,1);
   case DESCR(0,0,1,1,0,0,0,0, 1,0,0, 0, 1,0,1): /* 16  8 8  16 16 */
                                                 return BYTE(0,1,1,1,0,1,0,1);
   case DESCR(0,0,0,0,0,0,0,0, 1,0,1, 0, 1,0,1): /* 16  16   16 16 */
                                                 return BYTE(0,1,0,1,0,1,0,1);

   case DESCR(0,0,0,0,1,1,1,1, 0,1,0, 0, 0,0,0): /* 32  8 8 8 8 */
                                                 return BYTE(0,0,0,1,1,1,1,1);
   case DESCR(0,0,0,0,1,1,0,0, 0,1,0, 0, 0,0,1): /* 32  8 8 16  */
                                                 return BYTE(0,0,0,1,1,1,0,1);
   case DESCR(0,0,0,0,0,0,1,1, 0,1,0, 0, 1,0,0): /* 32  16  8 8 */
                                                 return BYTE(0,0,0,1,0,1,1,1);
   case DESCR(0,0,0,0,0,0,0,0, 0,1,0, 0, 1,0,1): /* 32  16  16  */
                                                 return BYTE(0,0,0,1,0,1,0,1);

   case DESCR(1,1,1,1,0,0,0,0, 0,0,0, 0, 0,1,0): /* 8 8 8 8  32 */
                                                 return BYTE(1,1,1,1,0,0,0,1);
   case DESCR(1,1,0,0,0,0,0,0, 0,0,1, 0, 0,1,0): /* 8 8 16   32 */
                                                 return BYTE(1,1,0,1,0,0,0,1);
   case DESCR(0,0,1,1,0,0,0,0, 1,0,0, 0, 0,1,0): /* 16  8 8  32 */
                                                 return BYTE(0,1,1,1,0,0,0,1);
   case DESCR(0,0,0,0,0,0,0,0, 1,0,1, 0, 0,1,0): /* 16  16   32 */
                                                 return BYTE(0,1,0,1,0,0,0,1);

   case DESCR(0,0,0,0,0,0,0,0, 0,1,0, 0, 0,1,0): /* 32 32 */
                                                 return BYTE(0,0,0,1,0,0,0,1);

   case DESCR(0,0,0,0,0,0,0,0, 0,0,0, 1, 0,0,0): /* 64 */
                                                 return BYTE(0,0,0,0,0,0,0,1);

   default: return BYTE(0,0,0,0,0,0,0,0); 
                   /* INVALID - any valid descr produces at least one
                      valid bit in tree[0..7]*/
   }
   /* NOTREACHED*/
   tl_assert(0);

#  undef DESCR
#  undef BYTE
}

__attribute__((unused))
static Bool is_sane_Descr ( UShort descr ) {
   return descr_to_validbits(descr) != 0;
}

static void sprintf_Descr ( /*OUT*/UChar* dst, UShort descr ) {
   VG_(sprintf)(dst, 
                "%d%d%d%d%d%d%d%d %d%d%d %d %d%d%d",
                (Int)((descr & TREE_DESCR_8_7) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_6) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_5) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_4) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_3) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_2) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_1) ? 1 : 0),
                (Int)((descr & TREE_DESCR_8_0) ? 1 : 0),
                (Int)((descr & TREE_DESCR_16_3) ? 1 : 0),
                (Int)((descr & TREE_DESCR_32_1) ? 1 : 0),
                (Int)((descr & TREE_DESCR_16_2) ? 1 : 0),
                (Int)((descr & TREE_DESCR_64)   ? 1 : 0),
                (Int)((descr & TREE_DESCR_16_1) ? 1 : 0),
                (Int)((descr & TREE_DESCR_32_0) ? 1 : 0),
                (Int)((descr & TREE_DESCR_16_0) ? 1 : 0)
   );
}
static void sprintf_Byte ( /*OUT*/UChar* dst, UChar byte ) {
   VG_(sprintf)(dst, "%d%d%d%d%d%d%d%d",
                     (Int)((byte & 128) ? 1 : 0),
                     (Int)((byte &  64) ? 1 : 0),
                     (Int)((byte &  32) ? 1 : 0),
                     (Int)((byte &  16) ? 1 : 0),
                     (Int)((byte &   8) ? 1 : 0),
                     (Int)((byte &   4) ? 1 : 0),
                     (Int)((byte &   2) ? 1 : 0),
                     (Int)((byte &   1) ? 1 : 0)
   );
}

static Bool is_sane_Descr_and_Tree ( UShort descr, UInt* tree ) {
   Word  i;
   UChar validbits = descr_to_validbits(descr);
   UChar buf[128], buf2[128];
   if (validbits == 0)
      goto bad;
   for (i = 0; i < 8; i++) {
      if (validbits & (1<<i)) {
         if (!is_SHVAL_valid(tree[i]))
            goto bad;
      } else {
         if (tree[i] != 0)
            goto bad;
      }
   }
   return True;
  bad:
   sprintf_Descr( buf, descr );
   sprintf_Byte( buf2, validbits );
   VG_(printf)("is_sane_Descr_and_Tree: bad tree {\n");
   VG_(printf)("   validbits 0x%02lx    %s\n", (UWord)validbits, buf2);
   VG_(printf)("       descr 0x%04lx  %s\n", (UWord)descr, buf);
   for (i = 0; i < 8; i++)
      VG_(printf)("   [%ld] 0x%08x\n", i, tree[i]);
   VG_(printf)("}\n");
   return 0;
}


static Bool is_sane_CacheLine ( CacheLine* cl )
{
   Word tno, cloff;

   if (!cl) goto bad;

   for (tno = 0, cloff = 0;  tno < N_LINE_TREES;  tno++, cloff += 8) {
      UShort descr = cl->descrs[tno];
      UInt*  tree  = &cl->svals[cloff];
      if (!is_sane_Descr_and_Tree(descr, tree))
         goto bad;
   }
   tl_assert(cloff == N_LINE_ARANGE);
   return True;
  bad:
   pp_CacheLine(cl);
   return False;
}


static UShort normalise_tree ( /*MOD*/UInt* tree ) {
   Word   i;
   UShort descr;
   /* pre: incoming tree[0..7] does not have any invalid shvals, in
      particular no zeroes. */
   for (i = 0; i < 8; i++)
      tl_assert(tree[i] != 0);
   
   descr = TREE_DESCR_8_7 | TREE_DESCR_8_6 | TREE_DESCR_8_5
           | TREE_DESCR_8_4 | TREE_DESCR_8_3 | TREE_DESCR_8_2
           | TREE_DESCR_8_1 | TREE_DESCR_8_0;
   /* build 16-bit layer */
   if (tree[1] == tree[0]) {
      tree[1] = 0/*INVALID*/;
      descr &= ~(TREE_DESCR_8_1 | TREE_DESCR_8_0);
      descr |= TREE_DESCR_16_0;
   }
   if (tree[3] == tree[2]) {
      tree[3] = 0/*INVALID*/;
      descr &= ~(TREE_DESCR_8_3 | TREE_DESCR_8_2);
      descr |= TREE_DESCR_16_1;
   }
   if (tree[5] == tree[4]) {
      tree[5] = 0/*INVALID*/;
      descr &= ~(TREE_DESCR_8_5 | TREE_DESCR_8_4);
      descr |= TREE_DESCR_16_2;
   }
   if (tree[7] == tree[6]) {
      tree[7] = 0/*INVALID*/;
      descr &= ~(TREE_DESCR_8_7 | TREE_DESCR_8_6);
      descr |= TREE_DESCR_16_3;
   }
   /* build 32-bit layer */
   if (tree[2] == tree[0]
       && (descr & TREE_DESCR_16_1) && (descr & TREE_DESCR_16_0)) {
      tree[2] = 0; /* [3,1] must already be 0 */
      descr &= ~(TREE_DESCR_16_1 | TREE_DESCR_16_0);
      descr |= TREE_DESCR_32_0;
   }
   if (tree[6] == tree[4]
       && (descr & TREE_DESCR_16_3) && (descr & TREE_DESCR_16_2)) {
      tree[6] = 0; /* [7,5] must already be 0 */
      descr &= ~(TREE_DESCR_16_3 | TREE_DESCR_16_2);
      descr |= TREE_DESCR_32_1;
   }
   /* build 64-bit layer */
   if (tree[4] == tree[0]
       && (descr & TREE_DESCR_32_1) && (descr & TREE_DESCR_32_0)) {
      tree[4] = 0; /* [7,6,5,3,2,1] must already be 0 */
      descr &= ~(TREE_DESCR_32_1 | TREE_DESCR_32_0);
      descr |= TREE_DESCR_64;
   }
   return descr;
}

/* This takes a cacheline where all the data is at the leaves
   (w8[..]) and builds a correctly normalised tree. */
static void normalise_CacheLine ( /*MOD*/CacheLine* cl )
{
   Word tno, cloff;
   for (tno = 0, cloff = 0;  tno < N_LINE_TREES;  tno++, cloff += 8) {
      UInt* tree = &cl->svals[cloff];
      cl->descrs[tno] = normalise_tree( tree );
   }
   tl_assert(cloff == N_LINE_ARANGE);
   if (SCE_CACHELINE)
      tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   stats__cline_normalises++;
}


static 
UInt* sequentialise_tree ( /*MOD*/UInt* dst, /*OUT*/Bool* anyShared,
                           UShort descr, UInt* tree ) {
   UInt* dst0 = dst;
   *anyShared = False;

#  define PUT(_n,_v)                                \
      do { Word i;                                  \
           if (is_SHVAL_Sh(_v))                     \
              *anyShared = True;                    \
           for (i = 0; i < (_n); i++)               \
                  *dst++ = (_v);                    \
      } while (0)

   /* byte 0 */
   if (descr & TREE_DESCR_64)   PUT(8, tree[0]); else
   if (descr & TREE_DESCR_32_0) PUT(4, tree[0]); else
   if (descr & TREE_DESCR_16_0) PUT(2, tree[0]); else
   if (descr & TREE_DESCR_8_0)  PUT(1, tree[0]);
   /* byte 1 */
   if (descr & TREE_DESCR_8_1)  PUT(1, tree[1]);
   /* byte 2 */
   if (descr & TREE_DESCR_16_1) PUT(2, tree[2]); else
   if (descr & TREE_DESCR_8_2)  PUT(1, tree[2]);
   /* byte 3 */
   if (descr & TREE_DESCR_8_3)  PUT(1, tree[3]);
   /* byte 4 */
   if (descr & TREE_DESCR_32_1) PUT(4, tree[4]); else
   if (descr & TREE_DESCR_16_2) PUT(2, tree[4]); else
   if (descr & TREE_DESCR_8_4)  PUT(1, tree[4]);
   /* byte 5 */
   if (descr & TREE_DESCR_8_5)  PUT(1, tree[5]);
   /* byte 6 */
   if (descr & TREE_DESCR_16_3) PUT(2, tree[6]); else
   if (descr & TREE_DESCR_8_6)  PUT(1, tree[6]);
   /* byte 7 */
   if (descr & TREE_DESCR_8_7)  PUT(1, tree[7]);

#  undef PUT

   tl_assert( (((Char*)dst) - ((Char*)dst0)) == 8 * sizeof(UInt) );
   return dst;
}

/* Write the cacheline 'wix' to backing store.  Where it ends up
   is determined by its tag field. */
static
Bool sequentialise_CacheLine ( /*OUT*/UInt* dst, Word nDst, CacheLine* src )
{
   Word  tno, cloff;
   Bool  anyShared = False;
   UInt* dst0      = dst;

   for (tno = 0, cloff = 0;  tno < N_LINE_TREES;  tno++, cloff += 8) {
      UShort descr = src->descrs[tno];
      UInt*  tree  = &src->svals[cloff];
      Bool   bTmp  = False;
      dst = sequentialise_tree ( dst, &bTmp, descr, tree );
      anyShared |= bTmp;
   }
   tl_assert(cloff == N_LINE_ARANGE);

   /* Assert we wrote N_LINE_ARANGE shadow values. */
   tl_assert( ((HChar*)dst) - ((HChar*)dst0) 
              == nDst * sizeof(UInt) );

   return anyShared;
}


static __attribute__((noinline)) void cacheline_wback ( UWord wix )
{
   Word        i, j;
   Bool        anyShared = False;
   Addr        tag;
   SecMap*     sm;
   CacheLine*  cl;
   CacheLineZ* lineZ;
   CacheLineF* lineF;
   Word        zix, fix;
   UInt        shvals[N_LINE_ARANGE];
   UInt        sv;

   if (0)
   VG_(printf)("scache wback line %d\n", (Int)wix);

   tl_assert(wix >= 0 && wix < N_WAY_NENT);

   tag =  cache_shmem.tags0[wix];
   cl  = &cache_shmem.lyns0[wix];

   /* The cache line may have been invalidated; if so, ignore it. */
   if (!is_valid_scache_tag(tag))
      return;

   /* Where are we going to put it? */
   sm         = NULL;
   lineZ      = NULL;
   lineF      = NULL;
   zix = fix = -1;

   find_Z_for_writing( &sm, &zix, tag );
   tl_assert(sm);
   tl_assert(zix >= 0 && zix < N_SECMAP_ZLINES);
   lineZ = &sm->linesZ[zix];

   /* Generate the data to be stored */
   if (SCE_CACHELINE)
      tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   anyShared = sequentialise_CacheLine( shvals, N_LINE_ARANGE, cl );

   lineZ->dict[0] = lineZ->dict[1] 
                  = lineZ->dict[2] = lineZ->dict[3] = 0;

   for (i = 0; i < N_LINE_ARANGE; i++) {

      sv = shvals[i];
      for (j = 0; j < 4; j++) {
         if (sv == lineZ->dict[j])
            goto dict_ok;
      }
      for (j = 0; j < 4; j++) {
         if (lineZ->dict[j] == 0)
            break;
      }
      tl_assert(j >= 0 && j <= 4);
      if (j == 4) break; /* we'll have to use the f rep */
      tl_assert(is_SHVAL_valid(sv));
      lineZ->dict[j] = sv;
     dict_ok:
      write_twobit_array( lineZ->ix2s, i, j );

   }

   tl_assert(i >= 0 && i <= N_LINE_ARANGE);

   if (i < N_LINE_ARANGE) {
      /* cannot use the compressed rep.  Use f rep instead. */
      alloc_F_for_writing( sm, &fix );
      tl_assert(sm->linesF);
      tl_assert(sm->linesF_size > 0);
      tl_assert(fix >= 0 && fix < sm->linesF_size);
      lineF = &sm->linesF[fix];
      tl_assert(!lineF->inUse);
      lineZ->dict[0] = lineZ->dict[2] = lineZ->dict[3] = 0;
      lineZ->dict[1] = (UInt)fix;
      lineF->inUse = True;
      for (i = 0; i < N_LINE_ARANGE; i++) {
         sv = shvals[i];
         tl_assert(is_SHVAL_valid(sv));
         lineF->w32s[i] = sv;
      }
      stats__cache_F_wbacks++;
   } else {
      stats__cache_Z_wbacks++;
   }

   if (anyShared)
      sm->mbHasShared = True;

   /* mb_tidy_one_cacheline(); */
}

/* Fetch the cacheline 'wix' from the backing store.  The tag
   associated with 'wix' is assumed to have already been filled in;
   hence that is used to determine where in the backing store to read
   from. */
static __attribute__((noinline)) void cacheline_fetch ( UWord wix )
{
   Word        i;
   Addr        tag;
   CacheLine*  cl;
   CacheLineZ* lineZ;
   CacheLineF* lineF;

   if (0)
   VG_(printf)("scache fetch line %d\n", (Int)wix);

   tl_assert(wix >= 0 && wix < N_WAY_NENT);

   tag =  cache_shmem.tags0[wix];
   cl  = &cache_shmem.lyns0[wix];

   /* reject nonsense requests */
   tl_assert(is_valid_scache_tag(tag));

   lineZ = NULL;
   lineF = NULL;
   find_ZF_for_reading( &lineZ, &lineF, tag );
   tl_assert( (lineZ && !lineF) || (!lineZ && lineF) );

   /* expand the data into the bottom layer of the tree, then get
      cacheline_normalise to build the descriptor array. */
   if (lineF) {
      tl_assert(lineF->inUse);
      for (i = 0; i < N_LINE_ARANGE; i++) {
         cl->svals[i] = lineF->w32s[i];
      }
      stats__cache_F_fetches++;
   } else {
      for (i = 0; i < N_LINE_ARANGE; i++) {
         UInt sv;
         UWord ix = read_twobit_array( lineZ->ix2s, i );
         tl_assert(ix >= 0 && ix <= 3);
         sv = lineZ->dict[ix];
         tl_assert(sv != 0);
         cl->svals[i] = sv;
      }
      stats__cache_Z_fetches++;
   }
   normalise_CacheLine( cl );
}

static void shmem__invalidate_scache ( void ) {
   Word wix;
   if (0) VG_(printf)("scache inval\n");
   tl_assert(!is_valid_scache_tag(1));
   for (wix = 0; wix < N_WAY_NENT; wix++) {
      cache_shmem.tags0[wix] = 1/*INVALID*/;
   }
   stats__cache_invals++;
}

static void shmem__flush_and_invalidate_scache ( void ) {
   Word wix;
   Addr tag;
   if (0) VG_(printf)("scache flush and invalidate\n");
   tl_assert(!is_valid_scache_tag(1));
   for (wix = 0; wix < N_WAY_NENT; wix++) {
      tag = cache_shmem.tags0[wix];
      if (tag == 1/*INVALID*/) {
         /* already invalid; nothing to do */
      } else {
         tl_assert(is_valid_scache_tag(tag));
         cacheline_wback( wix );
      }
      cache_shmem.tags0[wix] = 1/*INVALID*/;
   }
   stats__cache_flushes++;
   stats__cache_invals++;
}


/* ------------ Basic shadow memory read/write ops ------------ */

static inline Bool aligned16 ( Addr a ) {
   return 0 == (a & 1);
}
static inline Bool aligned32 ( Addr a ) {
   return 0 == (a & 3);
}
static inline Bool aligned64 ( Addr a ) {
   return 0 == (a & 7);
}
static inline UWord get_cacheline_offset ( Addr a ) {
   return (UWord)(a & (N_LINE_ARANGE - 1));
}
static inline UWord get_treeno ( Addr a ) {
   return get_cacheline_offset(a) >> 3;
}
static inline UWord get_tree_offset ( Addr a ) {
   return a & 7;
}

static __attribute__((noinline))
       CacheLine* get_cacheline_MISS ( Addr a ); /* fwds */
static inline CacheLine* get_cacheline ( Addr a )
{
   /* tag is 'a' with the in-line offset masked out, 
      eg a[31]..a[4] 0000 */
   Addr       tag = a & ~(N_LINE_ARANGE - 1);
   UWord      wix = (a >> N_LINE_BITS) & (N_WAY_NENT - 1);
   stats__cache_totrefs++;
   if (LIKELY(tag == cache_shmem.tags0[wix])) {
      return &cache_shmem.lyns0[wix];
   } else {
      return get_cacheline_MISS( a );
   }
}

static __attribute__((noinline))
       CacheLine* get_cacheline_MISS ( Addr a )
{
   /* tag is 'a' with the in-line offset masked out, 
      eg a[31]..a[4] 0000 */

   CacheLine* cl;
   Addr*      tag_old_p;
   Addr       tag = a & ~(N_LINE_ARANGE - 1);
   UWord      wix = (a >> N_LINE_BITS) & (N_WAY_NENT - 1);

   tl_assert(tag != cache_shmem.tags0[wix]);

   /* Dump the old line into the backing store. */
   stats__cache_totmisses++;

   cl        = &cache_shmem.lyns0[wix];
   tag_old_p = &cache_shmem.tags0[wix];

   if (is_valid_scache_tag( *tag_old_p )) {
      /* EXPENSIVE and REDUNDANT: callee does it */
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
      cacheline_wback( wix );
   }
   /* and reload the new one */
   *tag_old_p = tag;
   cacheline_fetch( wix );
   if (SCE_CACHELINE)
      tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   return cl;
}

static UShort pulldown_to_32 ( /*MOD*/UInt* tree, UWord toff, UShort descr ) {
   stats__cline_64to32pulldown++;
   switch (toff) {
      case 0: case 4:
         tl_assert(descr & TREE_DESCR_64);
         tree[4] = tree[0];
         descr &= ~TREE_DESCR_64;
         descr |= (TREE_DESCR_32_1 | TREE_DESCR_32_0);
         break;
      default:
         tl_assert(0);
   }
   return descr;
}

static UShort pulldown_to_16 ( /*MOD*/UInt* tree, UWord toff, UShort descr ) {
   stats__cline_32to16pulldown++;
   switch (toff) {
      case 0: case 2:
         if (!(descr & TREE_DESCR_32_0)) {
            descr = pulldown_to_32(tree, 0, descr);
         }
         tl_assert(descr & TREE_DESCR_32_0);
         tree[2] = tree[0];
         descr &= ~TREE_DESCR_32_0;
         descr |= (TREE_DESCR_16_1 | TREE_DESCR_16_0);
         break;
      case 4: case 6:
         if (!(descr & TREE_DESCR_32_1)) {
            descr = pulldown_to_32(tree, 4, descr);
         }
         tl_assert(descr & TREE_DESCR_32_1);
         tree[6] = tree[4];
         descr &= ~TREE_DESCR_32_1;
         descr |= (TREE_DESCR_16_3 | TREE_DESCR_16_2);
         break;
      default:
         tl_assert(0);
   }
   return descr;
}

static UShort pulldown_to_8 ( /*MOD*/UInt* tree, UWord toff, UShort descr ) {
   stats__cline_16to8pulldown++;
   switch (toff) {
      case 0: case 1:
         if (!(descr & TREE_DESCR_16_0)) {
            descr = pulldown_to_16(tree, 0, descr);
         }
         tl_assert(descr & TREE_DESCR_16_0);
         tree[1] = tree[0];
         descr &= ~TREE_DESCR_16_0;
         descr |= (TREE_DESCR_8_1 | TREE_DESCR_8_0);
         break;
      case 2: case 3:
         if (!(descr & TREE_DESCR_16_1)) {
            descr = pulldown_to_16(tree, 2, descr);
         }
         tl_assert(descr & TREE_DESCR_16_1);
         tree[3] = tree[2];
         descr &= ~TREE_DESCR_16_1;
         descr |= (TREE_DESCR_8_3 | TREE_DESCR_8_2);
         break;
      case 4: case 5:
         if (!(descr & TREE_DESCR_16_2)) {
            descr = pulldown_to_16(tree, 4, descr);
         }
         tl_assert(descr & TREE_DESCR_16_2);
         tree[5] = tree[4];
         descr &= ~TREE_DESCR_16_2;
         descr |= (TREE_DESCR_8_5 | TREE_DESCR_8_4);
         break;
      case 6: case 7:
         if (!(descr & TREE_DESCR_16_3)) {
            descr = pulldown_to_16(tree, 6, descr);
         }
         tl_assert(descr & TREE_DESCR_16_3);
         tree[7] = tree[6];
         descr &= ~TREE_DESCR_16_3;
         descr |= (TREE_DESCR_8_7 | TREE_DESCR_8_6);
         break;
      default:
         tl_assert(0);
   }
   return descr;
}


static UShort pullup_descr_to_16 ( UShort descr, UWord toff ) {
   UShort mask;
   switch (toff) {
      case 0:
         mask = TREE_DESCR_8_1 | TREE_DESCR_8_0;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_16_0;
         break;
      case 2:
         mask = TREE_DESCR_8_3 | TREE_DESCR_8_2;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_16_1;
         break;
      case 4:
         mask = TREE_DESCR_8_5 | TREE_DESCR_8_4;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_16_2;
         break;
      case 6:
         mask = TREE_DESCR_8_7 | TREE_DESCR_8_6;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_16_3;
         break;
      default:
         tl_assert(0);
   }
   return descr;
}

static UShort pullup_descr_to_32 ( UShort descr, UWord toff ) {
   UShort mask;
   switch (toff) {
      case 0:
         if (!(descr & TREE_DESCR_16_0))
            descr = pullup_descr_to_16(descr, 0);
         if (!(descr & TREE_DESCR_16_1))
            descr = pullup_descr_to_16(descr, 2);
         mask = TREE_DESCR_16_1 | TREE_DESCR_16_0;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_32_0;
         break;
      case 4:
         if (!(descr & TREE_DESCR_16_2))
            descr = pullup_descr_to_16(descr, 4);
         if (!(descr & TREE_DESCR_16_3))
            descr = pullup_descr_to_16(descr, 6);
         mask = TREE_DESCR_16_3 | TREE_DESCR_16_2;
         tl_assert( (descr & mask) == mask );
         descr &= ~mask;
         descr |= TREE_DESCR_32_1;
         break;
      default:
         tl_assert(0);
   }
   return descr;
}

static Bool valid_value_is_above_me_32 ( UShort descr, UWord toff ) {
   switch (toff) {
      case 0: case 4:
         return 0 != (descr & TREE_DESCR_64);
      default:
         tl_assert(0);
   }
}

static Bool valid_value_is_below_me_16 ( UShort descr, UWord toff ) {
   switch (toff) {
      case 0:
         return 0 != (descr & (TREE_DESCR_8_1 | TREE_DESCR_8_0));
      case 2:
         return 0 != (descr & (TREE_DESCR_8_3 | TREE_DESCR_8_2));
      case 4:
         return 0 != (descr & (TREE_DESCR_8_5 | TREE_DESCR_8_4));
      case 6:
         return 0 != (descr & (TREE_DESCR_8_7 | TREE_DESCR_8_6));
      default:
         tl_assert(0);
   }
}

static void shadow_mem_read8 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_read8s++;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 .. 7 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_8_0 << toff)) )) {
      UInt* tree = &cl->svals[tno << 3];
      cl->descrs[tno] = pulldown_to_8(tree, toff, descr);
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_read( thr_acc, a, svOld, 1 );
   cl->svals[cloff] = svNew;
}
static void shadow_mem_read16 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_read16s++;
   if (UNLIKELY(!aligned16(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0, 2, 4 or 6 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_16_0 << toff)) )) {
      if (valid_value_is_below_me_16(descr, toff)) {
         goto slowcase;
      } else {
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_16(tree, toff, descr);
      }
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_read( thr_acc, a, svOld, 2 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_16to8splits++;
   shadow_mem_read8( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_read8( thr_acc, a + 1, 0/*unused*/ );
}

__attribute__((noinline))
static void shadow_mem_read32_SLOW ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   if (UNLIKELY(!aligned32(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 or 4 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_32_0 << toff)) )) {
      if (valid_value_is_above_me_32(descr, toff)) {
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_32(tree, toff, descr);
      } else {
         goto slowcase;
      }
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_read( thr_acc, a, svOld, 4 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_32to16splits++;
   shadow_mem_read16( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_read16( thr_acc, a + 2, 0/*unused*/ );
}
inline
static void shadow_mem_read32 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_read32s++;
   if (UNLIKELY(!aligned32(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 or 4 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_32_0 << toff)) )) goto slowcase;
   { UInt* p = &cl->svals[cloff];
     *p = msm__handle_read( thr_acc, a, *p, 4 );
   }
   return;
  slowcase: /* misaligned, or not at this level in the tree */
   shadow_mem_read32_SLOW( thr_acc, a, uuOpaque );
}

inline
static void shadow_mem_read64 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_read64s++;
   if (UNLIKELY(!aligned64(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0, unused */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & TREE_DESCR_64) )) {
      goto slowcase;
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_read( thr_acc, a, svOld, 8 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_64to32splits++;
   shadow_mem_read32( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_read32( thr_acc, a + 4, 0/*unused*/ );
}

static void shadow_mem_write8 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_write8s++;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 .. 7 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_8_0 << toff)) )) {
      UInt* tree = &cl->svals[tno << 3];
      cl->descrs[tno] = pulldown_to_8(tree, toff, descr);
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_write( thr_acc, a, svOld, 1 );
   cl->svals[cloff] = svNew;
}
static void shadow_mem_write16 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_write16s++;
   if (UNLIKELY(!aligned16(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0, 2, 4 or 6 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_16_0 << toff)) )) {
      if (valid_value_is_below_me_16(descr, toff)) {
         goto slowcase;
      } else {
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_16(tree, toff, descr);
      }
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_write( thr_acc, a, svOld, 2 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_16to8splits++;
   shadow_mem_write8( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_write8( thr_acc, a + 1, 0/*unused*/ );
}

__attribute__((noinline))
static void shadow_mem_write32_SLOW ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   if (UNLIKELY(!aligned32(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 or 4 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_32_0 << toff)) )) {
      if (valid_value_is_above_me_32(descr, toff)) {
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_32(tree, toff, descr);
      } else {
         goto slowcase;
      }
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_write( thr_acc, a, svOld, 4 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_32to16splits++;
   shadow_mem_write16( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_write16( thr_acc, a + 2, 0/*unused*/ );
}
inline
static void shadow_mem_write32 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_write32s++;
   if (UNLIKELY(!aligned32(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 or 4 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_32_0 << toff)) )) goto slowcase;
   { UInt* p = &cl->svals[cloff];
     *p = msm__handle_write( thr_acc, a, *p, 4 );
   }
   return;
  slowcase: /* misaligned, or must go further down the tree */
   shadow_mem_write32_SLOW( thr_acc, a, uuOpaque );
}

inline
static void shadow_mem_write64 ( Thread* thr_acc, Addr a, UInt uuOpaque ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UInt       svOld, svNew;
   UShort     descr;
   stats__cline_write64s++;
   if (UNLIKELY(!aligned64(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0, unused */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & TREE_DESCR_64) )) {
      goto slowcase;
   }
   svOld = cl->svals[cloff];
   svNew = msm__handle_write( thr_acc, a, svOld, 8 );
   cl->svals[cloff] = svNew;
   return;
  slowcase: /* misaligned, or must go further down the tree */
   stats__cline_64to32splits++;
   shadow_mem_write32( thr_acc, a + 0, 0/*unused*/ );
   shadow_mem_write32( thr_acc, a + 4, 0/*unused*/ );
}

static void shadow_mem_set8 ( Thread* uu_thr_acc, Addr a, UInt svNew ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_set8s++;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 .. 7 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_8_0 << toff)) )) {
      UInt* tree = &cl->svals[tno << 3];
      cl->descrs[tno] = pulldown_to_8(tree, toff, descr);
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
   }
   cl->svals[cloff] = svNew;
}
static void shadow_mem_set16 ( Thread* uu_thr_acc, Addr a, UInt svNew ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_set16s++;
   if (UNLIKELY(!aligned16(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0, 2, 4 or 6 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_16_0 << toff)) )) {
      if (valid_value_is_below_me_16(descr, toff)) {
         /* Writing at this level.  Need to fix up 'descr'. */
         cl->descrs[tno] = pullup_descr_to_16(descr, toff);
         /* At this point, the tree does not match cl->descr[tno] any
            more.  The assignments below will fix it up. */
      } else {
         /* We can't indiscriminately write on the w16 node as in the
            w64 case, as that might make the node inconsistent with
            its parent.  So first, pull down to this level. */
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_16(tree, toff, descr);
      if (SCE_CACHELINE)
         tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
      }
   }
   cl->svals[cloff + 0] = svNew;
   cl->svals[cloff + 1] = 0;
   return;
  slowcase: /* misaligned */
   stats__cline_16to8splits++;
   shadow_mem_set8( uu_thr_acc, a + 0, svNew );
   shadow_mem_set8( uu_thr_acc, a + 1, svNew );
}
static void shadow_mem_set32 ( Thread* uu_thr_acc, Addr a, UInt svNew ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_set32s++;
   if (UNLIKELY(!aligned32(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 or 4 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_32_0 << toff)) )) {
      if (valid_value_is_above_me_32(descr, toff)) {
         /* We can't indiscriminately write on the w32 node as in the
            w64 case, as that might make the node inconsistent with
            its parent.  So first, pull down to this level. */
         UInt* tree = &cl->svals[tno << 3];
         cl->descrs[tno] = pulldown_to_32(tree, toff, descr);
         if (SCE_CACHELINE)
            tl_assert(is_sane_CacheLine(cl)); /* EXPENSIVE */
      } else {
         /* Writing at this level.  Need to fix up 'descr'. */
         cl->descrs[tno] = pullup_descr_to_32(descr, toff);
         /* At this point, the tree does not match cl->descr[tno] any
            more.  The assignments below will fix it up. */
      }
   }
   cl->svals[cloff + 0] = svNew;
   cl->svals[cloff + 1] = 0;
   cl->svals[cloff + 2] = 0;
   cl->svals[cloff + 3] = 0;
   return;
  slowcase: /* misaligned */
   stats__cline_32to16splits++;
   shadow_mem_set16( uu_thr_acc, a + 0, svNew );
   shadow_mem_set16( uu_thr_acc, a + 2, svNew );
}
inline
static void shadow_mem_set64 ( Thread* uu_thr_acc, Addr a, UInt svNew ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   stats__cline_set64s++;
   if (UNLIKELY(!aligned64(a))) goto slowcase;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 */
   cl->descrs[tno] = TREE_DESCR_64;
   cl->svals[cloff + 0] = svNew;
   cl->svals[cloff + 1] = 0;
   cl->svals[cloff + 2] = 0;
   cl->svals[cloff + 3] = 0;
   cl->svals[cloff + 4] = 0;
   cl->svals[cloff + 5] = 0;
   cl->svals[cloff + 6] = 0;
   cl->svals[cloff + 7] = 0;
   return;
  slowcase: /* misaligned */
   stats__cline_64to32splits++;
   shadow_mem_set32( uu_thr_acc, a + 0, svNew );
   shadow_mem_set32( uu_thr_acc, a + 4, svNew );
}

static UInt shadow_mem_get8 ( Addr a ) {
   CacheLine* cl; 
   UWord      cloff, tno, toff;
   UShort     descr;
   stats__cline_get8s++;
   cl    = get_cacheline(a);
   cloff = get_cacheline_offset(a);
   tno   = get_treeno(a);
   toff  = get_tree_offset(a); /* == 0 .. 7 */
   descr = cl->descrs[tno];
   if (UNLIKELY( !(descr & (TREE_DESCR_8_0 << toff)) )) {
      UInt* tree = &cl->svals[tno << 3];
      cl->descrs[tno] = pulldown_to_8(tree, toff, descr);
   }
   return cl->svals[cloff];
}

static void shadow_mem_copy8 ( Addr src, Addr dst, Bool normalise ) {
   UInt       sv;
   stats__cline_copy8s++;
   sv = shadow_mem_get8( src );

   if (UNLIKELY(clo_trace_level > 0)) {
      if (dst == clo_trace_addr) {
         Thread* thr    = get_current_Thread();
         UInt    sv_old = shadow_mem_get8( dst );
         msm__show_state_change( thr, dst, 1, 'w', sv_old, sv );
      }
   }

   shadow_mem_set8( NULL/*unused*/, dst, sv );
}


/* ------------ Shadow memory range setting ops ------------ */

static void shadow_mem_modify_range(
               Thread* thr, 
               Addr    a, 
               SizeT   len,
               void    (*fn8) (Thread*,Addr,UInt),
               void    (*fn16)(Thread*,Addr,UInt),
               void    (*fn32)(Thread*,Addr,UInt),
               void    (*fn64)(Thread*,Addr,UInt),
               UInt    opaque
            )
{
   /* fast track a couple of common cases */
   if (len == 4 && aligned32(a)) {
      fn32( thr, a, opaque );
      return;
   }
   if (len == 8 && aligned64(a)) {
      fn64( thr, a, opaque );
      return;
   }

   /* be completely general (but as efficient as possible) */
   if (len == 0) return;

   if (!aligned16(a) && len >= 1) {
      fn8( thr, a, opaque );
      a += 1;
      len -= 1;
      tl_assert(aligned16(a));
   }
   if (len == 0) return;

   if (!aligned32(a) && len >= 2) {
      fn16( thr, a, opaque );
      a += 2;
      len -= 2;
      tl_assert(aligned32(a));
   }
   if (len == 0) return;

   if (!aligned64(a) && len >= 4) {
      fn32( thr, a, opaque );
      a += 4;
      len -= 4;
      tl_assert(aligned64(a));
   }
   if (len == 0) return;

   if (len >= 8) {
      tl_assert(aligned64(a));
      while (len >= 8) {
         fn64( thr, a, opaque );
         a += 8;
         len -= 8;
      }
      tl_assert(aligned64(a));
   }
   if (len == 0) return;

   if (len >= 4)
      tl_assert(aligned32(a));
   if (len >= 4) {
      fn32( thr, a, opaque );
      a += 4;
      len -= 4;
   }
   if (len == 0) return;

   if (len >= 2)
      tl_assert(aligned16(a));
   if (len >= 2) {
      fn16( thr, a, opaque );
      a += 2;
      len -= 2;
   }
   if (len == 0) return;

   if (len >= 1) {
      fn8( thr, a, opaque );
      a += 1;
      len -= 1;
   }
   tl_assert(len == 0);
}

/* Block-copy states (needed for implementing realloc()). */
static void shadow_mem_copy_range ( Addr src, Addr dst, SizeT len )
{
   SizeT i;
   if (len == 0)
      return;
   /* To be simple, just copy byte by byte.  But so as not to wreck
      performance for later accesses to dst[0 .. len-1], normalise
      destination lines as we finish with them, and also normalise the
      line containing the first and last address. */
   for (i = 0; i < len; i++) {
      Bool normalise
         = get_cacheline_offset( dst+i+1 ) == 0 /* last in line */
           || i == 0       /* first in range */
           || i == len-1;  /* last in range */
      shadow_mem_copy8( src+i, dst+i, normalise );
   }
}

static void shadow_mem_read_range ( Thread* thr, Addr a, SizeT len ) {
   shadow_mem_modify_range( thr, a, len, 
                            shadow_mem_read8,
                            shadow_mem_read16,
                            shadow_mem_read32,
                            shadow_mem_read64,
                            0/*opaque,ignored*/ );
}

static void shadow_mem_write_range ( Thread* thr, Addr a, SizeT len ) {
   shadow_mem_modify_range( thr, a, len, 
                            shadow_mem_write8,
                            shadow_mem_write16,
                            shadow_mem_write32,
                            shadow_mem_write64,
                            0/*opaque,ignored*/ );
}

static void shadow_mem_make_New ( Thread* thr, Addr a, SizeT len )
{
   if (UNLIKELY(clo_trace_level > 0)) {
      if (len > 0 && a <= clo_trace_addr && clo_trace_addr < a+len) {
         UInt sv_old = shadow_mem_get8( clo_trace_addr );
         msm__show_state_change( thr, a, (Int)len, 'p', sv_old, SHVAL_New );
      }
   }
   shadow_mem_modify_range( thr, a, len, 
                            shadow_mem_set8,
                            shadow_mem_set16,
                            shadow_mem_set32,
                            shadow_mem_set64,
                            SHVAL_New/*opaque*/ );
}


/* Putting memory into the NoAccess state.  This is hugely complicated
   by the problem of memory that contains locks.

   1. Examine the .mbHasLocks fields in all SecMaps in the range to be
      deleted.  This quickly indicates if there are or might be any
      locks in the range to be deleted.  Note that .mbHasLocks fields on
      SecMaps are not subject to scaching, so it safe to look at them
      without flushing the scache.

   2. Set the range to NoAccess.  Clear the .mbHasShared and
      .mbHasLocks hint bits for any completely vacated SecMaps.
      Clearing the hint bits isn't necessary for correctness, but it
      is important to avoid ending up with hint bits being permanently
      set, which would render them pointless.

   3. If (1) indicated "definitely no locks", we're done.  This is
      the fast and hopefully common case.

   Otherwise, the range contains some locks (or may do), so we have to
   go to considerable effort to tidy up.

   4. Make up a set containing the locks which are deleted:

      ToDelete = NULL

      for each lk in map_locks {
         if lk's guest addr falls in the range to memory be deleted
            add lk to ToDelete

         if lk is held, issue an error message - freeing memory
            containing a held lock
      }

   5. If ToDelete is empty, there were in fact no locks in the range,
      despite what the .mbHasLocks hint bits indicated.  We're done.

   6. Flush the scache.  This is necessary both to bring the SecMap
      .mbHasShared fields up to date, and to bring the actual shadow
      values up to date.  We will need to examine both of these.

      Invalidate the scache.  This is necessary because we will be
      modifying values in the backing store (SecMaps) and need
      subsequent shmem accesses to get the new values.

   7. Modify all shadow words, by removing ToDelete from the lockset
      of all ShM and ShR states.  Note this involves a complete scan
      over map_shmem, which is very expensive according to OProfile.
      Hence it depends critically on the size of each entry in
      map_shmem.  See comments on definition of N_SECMAP_BITS above.

      Why is it safe to do (7) after (2) ?  Because we're not
      interested in messing with ShR/M states which are going to be
      set to NoAccess anyway.

      Optimisation 1 (implemented): skip this step for SecMaps which
      do not have .mbHasShared set

      Optimisation 2 (not implemented): for each SecMap, have a
      summary lock set which is the union of all locks mentioned in
      locksets on this page (or any superset of it).  Then skip step
      (2) if the summary lockset does not intersect with ToDelete.

      That's potentially cheap, since the usual lockset refinement
      only shrinks locksets; hence there is no point in updating the
      summary lockset for ShM/R -> ShM/R transitions.  Therefore only
      need to do this for Excl->ShM/R transitions.

   8. Tell laog that these locks have disappeared.
*/
static void shadow_mem_make_NoAccess ( Thread* thr, Addr aIN, SizeT len )
{
   Lock*     lk;
   Addr      gla, sma, firstSM, lastSM, firstA, lastA;
   WordSetID locksToDelete;
   Bool      mbHasLocks;

   if (0 && len > 500)
      VG_(printf)("make NoAccess ( %p, %d )\n", aIN, len );

   if (len == 0) 
      return;

   /* --- Step 1 --- */

   firstA  = aIN;
   lastA   = aIN + len - 1;

   firstSM = shmem__round_to_SecMap_base( firstA );
   lastSM  = shmem__round_to_SecMap_base( lastA );
   tl_assert(firstSM <= lastSM);

   mbHasLocks = False;
   for (sma = firstSM; sma <= lastSM; sma += N_SECMAP_ARANGE) {
      if (shmem__get_mbHasLocks(sma)) {
         mbHasLocks = True;
         break;
      }
   }

   /* --- Step 2 --- */

   if (UNLIKELY(clo_trace_level > 0)) {
      if (len > 0 && firstA <= clo_trace_addr && clo_trace_addr <= lastA) {
         UInt sv_old = shadow_mem_get8( clo_trace_addr );
         msm__show_state_change( thr, firstA, (Int)len, 'p',
                                      sv_old, SHVAL_NoAccess );
      }
   }
   shadow_mem_modify_range( thr, firstA, len, 
                            shadow_mem_set8,
                            shadow_mem_set16,
                            shadow_mem_set32,
                            shadow_mem_set64,
                            SHVAL_NoAccess/*opaque*/ );

   for (sma = firstSM; sma <= lastSM; sma += N_SECMAP_ARANGE) {
      /* Is this sm entirely within the deleted range? */
      if (firstA <= sma && sma + N_SECMAP_ARANGE - 1 <= lastA) {
         /* Yes.  Clear the hint bits. */
         shmem__set_mbHasLocks( sma, False );
         shmem__set_mbHasShared( sma, False );
      }
   }

   /* --- Step 3 --- */

   if (!mbHasLocks)
      return;

   /* --- Step 4 --- */

   if (0) 
   VG_(printf)("shadow_mem_make_NoAccess(%p, %u, %p): maybe slow case\n",
               (void*)firstA, (UWord)len, (void*)lastA);
   locksToDelete = HG_(emptyWS)( univ_lsets );
   
   /* FIXME: don't iterate over the complete lock set */
   HG_(initIterFM)( map_locks );
   while (HG_(nextIterFM)( map_locks,
                           (Word*)(void*)&gla, (Word*)(void*)&lk )) {
      tl_assert(is_sane_LockN(lk));
      if (gla < firstA || gla > lastA)
         continue;
      locksToDelete = HG_(addToWS)( univ_lsets, locksToDelete, (Word)lk );
      /* If the lock is held, we must remove it from the currlock sets
         of all threads that hold it.  Also take the opportunity to
         report an error.  To report an error we need to know at least
         one of the threads that holds it; really we should mention
         them all, but that's too much hassle.  So choose one
         arbitrarily. */
      if (lk->heldBy) {
         tl_assert(!HG_(isEmptyBag)(lk->heldBy));
         record_error_FreeMemLock( (Thread*)HG_(anyElementOfBag)(lk->heldBy),
                                   lk );
         /* remove lock from locksets of all owning threads */
         remove_Lock_from_locksets_of_all_owning_Threads( lk );
         /* Leave lk->heldBy in place; del_Lock below will free it up. */
      }
   }
   HG_(doneIterFM)( map_locks );

   /* --- Step 5 --- */

   if (HG_(isEmptyWS)( univ_lsets, locksToDelete ))
      return;

   /* --- Step 6 --- */

   shmem__flush_and_invalidate_scache();

   /* --- Step 7 --- */

   if (0) 
   VG_(printf)("shadow_mem_make_NoAccess(%p, %u, %p): definitely slow case\n",
               (void*)firstA, (UWord)len, (void*)lastA);

   /* Modify all shadow words, by removing locksToDelete from the lockset
      of all ShM and ShR states.
      Optimisation 1: skip SecMaps which do not have .mbHasShared set
   */
   { Int        stats_SMs = 0, stats_SMs_scanned = 0;
     Addr       ga;
     SecMap*    sm;
     SecMapIter itr;
     UInt*      w32p;

     HG_(initIterFM)( map_shmem );
     while (HG_(nextIterFM)( map_shmem,
                             (Word*)(void*)&ga, (Word*)(void*)&sm )) {
        tl_assert(sm);
        stats_SMs++;
        /* Skip this SecMap if the summary bit indicates it is safe to
           do so. */
        if (!sm->mbHasShared)
           continue;
        stats_SMs_scanned++;
        initSecMapIter( &itr );
        while (stepSecMapIter( &w32p, &itr, sm )) {
           Bool isM;
           UInt wold, wnew, lset_old, tset_old, lset_new;
           wold = *w32p;
           if (LIKELY( !is_SHVAL_Sh(wold) ))
              continue;
           isM      = is_SHVAL_ShM(wold);
           lset_old = un_SHVAL_Sh_lset(wold);
           tset_old = un_SHVAL_Sh_tset(wold);
           lset_new = HG_(minusWS)( univ_lsets, lset_old, locksToDelete );
           wnew     = isM ? mk_SHVAL_ShM(tset_old, lset_new)
                          : mk_SHVAL_ShR(tset_old, lset_new);
           if (wnew != wold)
              *w32p = wnew;
        }
     }
     HG_(doneIterFM)( map_shmem );
     if (SHOW_EXPENSIVE_STUFF)
        VG_(printf)("shadow_mem_make_NoAccess: %d SMs, %d scanned\n", 
                    stats_SMs, stats_SMs_scanned);
   }

   /* Now we have to free up the Locks in locksToDelete and remove
      any mention of them from admin_locks and map_locks.  This is
      inefficient. */
   { Lock* lkprev = NULL;
     lk = admin_locks;
     while (True) {
        if (lk == NULL) break;
        if (lkprev) tl_assert(lkprev->admin == lk);

        if (!HG_(elemWS)(univ_lsets, locksToDelete, (Word)lk)) {
           lkprev = lk;
           lk = lk->admin;
           continue;
        }
        /* Need to delete 'lk' */
        if (lkprev == NULL) {
           admin_locks = lk->admin;
        } else {
           lkprev->admin = lk->admin;
        }
        /* and get it out of map_locks */
        map_locks_delete(lk->guestaddr);
        /* release storage (incl. associated .heldBy Bag) */
        { Lock* tmp = lk->admin;
          del_LockN(lk);
          lk = tmp;
        }
     }
   }

   /* --- Step 8 --- */

   /* update lock order acquisition graph */
   laog__handle_lock_deletions( locksToDelete );

   if (0) all__sanity_check("Make NoAccess");
}


/*----------------------------------------------------------------*/
/*--- Event handlers (evh__* functions)                        ---*/
/*--- plus helpers (evhH__* functions)                         ---*/
/*----------------------------------------------------------------*/

/*--------- Event handler helpers (evhH__* functions) ---------*/

/* Create a new segment for 'thr', making it depend (.prev) on its
   existing segment, bind together the SegmentID and Segment, and
   return both of them.  Also update 'thr' so it references the new
   Segment. */
static 
void evhH__start_new_segment_for_thread ( /*OUT*/SegmentID* new_segidP,
                                          /*OUT*/Segment** new_segP,
                                          Thread* thr )
{
   Segment* cur_seg;
   tl_assert(new_segP);
   tl_assert(new_segidP);
   tl_assert(is_sane_Thread(thr));
   cur_seg = map_segments_lookup( thr->csegid );
   tl_assert(cur_seg);
   tl_assert(cur_seg->thr == thr); /* all sane segs should point back
                                      at their owner thread. */
   *new_segP = mk_Segment( thr, cur_seg, NULL/*other*/ );
   *new_segidP = alloc_SegmentID();
   map_segments_add( *new_segidP, *new_segP );
   thr->csegid = *new_segidP;
}


/* The lock at 'lock_ga' has acquired a writer.  Make all necessary
   updates, and also do all possible error checks. */
static 
void evhH__post_thread_w_acquires_lock ( Thread* thr, 
                                         LockKind lkk, Addr lock_ga )
{
   Lock* lk; 

   /* Basically what we need to do is call lockN_acquire_writer.
      However, that will barf if any 'invalid' lock states would
      result.  Therefore check before calling.  Side effect is that
      'is_sane_LockN(lk)' is both a pre- and post-condition of this
      routine. 

      Because this routine is only called after successful lock
      acquisition, we should not be asked to move the lock into any
      invalid states.  Requests to do so are bugs in libpthread, since
      that should have rejected any such requests. */

   /* be paranoid w.r.t hint bits, even if lock_ga is complete
      nonsense */
   shmem__set_mbHasLocks( lock_ga, True );

   tl_assert(is_sane_Thread(thr));
   /* Try to find the lock.  If we can't, then create a new one with
      kind 'lkk'. */
   lk = map_locks_lookup_or_create( 
           lkk, lock_ga, map_threads_reverse_lookup_SLOW(thr) );
   tl_assert( is_sane_LockN(lk) );
   shmem__set_mbHasLocks( lock_ga, True );

   if (lk->heldBy == NULL) {
      /* the lock isn't held.  Simple. */
      tl_assert(!lk->heldW);
      lockN_acquire_writer( lk, thr );
      goto noerror;
   }

   /* So the lock is already held.  If held as a r-lock then
      libpthread must be buggy. */
   tl_assert(lk->heldBy);
   if (!lk->heldW) {
      record_error_Misc( thr, "Bug in libpthread: write lock "
                              "granted on rwlock which is currently rd-held");
      goto error;
   }

   /* So the lock is held in w-mode.  If it's held by some other
      thread, then libpthread must be buggy. */
   tl_assert(HG_(sizeUniqueBag)(lk->heldBy) == 1); /* from precondition */

   if (thr != (Thread*)HG_(anyElementOfBag)(lk->heldBy)) {
      record_error_Misc( thr, "Bug in libpthread: write lock "
                              "granted on mutex/rwlock which is currently "
                              "wr-held by a different thread");
      goto error;
   }

   /* So the lock is already held in w-mode by 'thr'.  That means this
      is an attempt to lock it recursively, which is only allowable
      for LK_mbRec kinded locks.  Since this routine is called only
      once the lock has been acquired, this must also be a libpthread
      bug. */
   if (lk->kind != LK_mbRec) {
      record_error_Misc( thr, "Bug in libpthread: recursive write lock "
                              "granted on mutex/wrlock which does not "
                              "support recursion");
      goto error;
   }

   /* So we are recursively re-locking a lock we already w-hold. */
   lockN_acquire_writer( lk, thr );
   goto noerror;

  noerror:
   /* check lock order acquisition graph, and update.  This has to
      happen before the lock is added to the thread's locksetA/W. */
   laog__pre_thread_acquires_lock( thr, lk );
   /* update the thread's held-locks set */
   thr->locksetA = HG_(addToWS)( univ_lsets, thr->locksetA, (Word)lk );
   thr->locksetW = HG_(addToWS)( univ_lsets, thr->locksetW, (Word)lk );
   /* fall through */

  error:
   tl_assert(is_sane_LockN(lk));
}


/* The lock at 'lock_ga' has acquired a reader.  Make all necessary
   updates, and also do all possible error checks. */
static 
void evhH__post_thread_r_acquires_lock ( Thread* thr, 
                                         LockKind lkk, Addr lock_ga )
{
   Lock* lk; 

   /* Basically what we need to do is call lockN_acquire_reader.
      However, that will barf if any 'invalid' lock states would
      result.  Therefore check before calling.  Side effect is that
      'is_sane_LockN(lk)' is both a pre- and post-condition of this
      routine. 

      Because this routine is only called after successful lock
      acquisition, we should not be asked to move the lock into any
      invalid states.  Requests to do so are bugs in libpthread, since
      that should have rejected any such requests. */

   /* be paranoid w.r.t hint bits, even if lock_ga is complete
      nonsense */
   shmem__set_mbHasLocks( lock_ga, True );

   tl_assert(is_sane_Thread(thr));
   /* Try to find the lock.  If we can't, then create a new one with
      kind 'lkk'.  Only a reader-writer lock can be read-locked,
      hence the first assertion. */
   tl_assert(lkk == LK_rdwr);
   lk = map_locks_lookup_or_create( 
           lkk, lock_ga, map_threads_reverse_lookup_SLOW(thr) );
   tl_assert( is_sane_LockN(lk) );
   shmem__set_mbHasLocks( lock_ga, True );

   if (lk->heldBy == NULL) {
      /* the lock isn't held.  Simple. */
      tl_assert(!lk->heldW);
      lockN_acquire_reader( lk, thr );
      goto noerror;
   }

   /* So the lock is already held.  If held as a w-lock then
      libpthread must be buggy. */
   tl_assert(lk->heldBy);
   if (lk->heldW) {
      record_error_Misc( thr, "Bug in libpthread: read lock "
                              "granted on rwlock which is "
                              "currently wr-held");
      goto error;
   }

   /* Easy enough.  In short anybody can get a read-lock on a rwlock
      provided it is either unlocked or already in rd-held. */
   lockN_acquire_reader( lk, thr );
   goto noerror;

  noerror:
   /* check lock order acquisition graph, and update.  This has to
      happen before the lock is added to the thread's locksetA/W. */
   laog__pre_thread_acquires_lock( thr, lk );
   /* update the thread's held-locks set */
   thr->locksetA = HG_(addToWS)( univ_lsets, thr->locksetA, (Word)lk );
   /* but don't update thr->locksetW, since lk is only rd-held */
   /* fall through */

  error:
   tl_assert(is_sane_LockN(lk));
}


/* The lock at 'lock_ga' is just about to be unlocked.  Make all
   necessary updates, and also do all possible error checks. */
static 
void evhH__pre_thread_releases_lock ( Thread* thr,
                                      Addr lock_ga, Bool isRDWR )
{
   Lock* lock;
   Word  n;

   /* This routine is called prior to a lock release, before
      libpthread has had a chance to validate the call.  Hence we need
      to detect and reject any attempts to move the lock into an
      invalid state.  Such attempts are bugs in the client.

      isRDWR is True if we know from the wrapper context that lock_ga
      should refer to a reader-writer lock, and is False if [ditto]
      lock_ga should refer to a standard mutex. */

   /* be paranoid w.r.t hint bits, even if lock_ga is complete
      nonsense */
   shmem__set_mbHasLocks( lock_ga, True );

   tl_assert(is_sane_Thread(thr));
   lock = map_locks_maybe_lookup( lock_ga );

   if (!lock) {
      /* We know nothing about a lock at 'lock_ga'.  Nevertheless
         the client is trying to unlock it.  So complain, then ignore
         the attempt. */
      record_error_UnlockBogus( thr, lock_ga );
      return;
   }

   tl_assert(lock->guestaddr == lock_ga);
   tl_assert(is_sane_LockN(lock));

   if (isRDWR && lock->kind != LK_rdwr) {
      record_error_Misc( thr, "pthread_rwlock_unlock with a "
                              "pthread_mutex_t* argument " );
   }
   if ((!isRDWR) && lock->kind == LK_rdwr) {
      record_error_Misc( thr, "pthread_mutex_unlock with a "
                              "pthread_rwlock_t* argument " );
   }

   if (!lock->heldBy) {
      /* The lock is not held.  This indicates a serious bug in the
         client. */
      tl_assert(!lock->heldW);
      record_error_UnlockUnlocked( thr, lock );
      tl_assert(!HG_(elemWS)( univ_lsets, thr->locksetA, (Word)lock ));
      tl_assert(!HG_(elemWS)( univ_lsets, thr->locksetW, (Word)lock ));
      goto error;
   }

   /* The lock is held.  Is this thread one of the holders?  If not,
      report a bug in the client. */
   n = HG_(elemBag)( lock->heldBy, (Word)thr );
   tl_assert(n >= 0);
   if (n == 0) {
      /* We are not a current holder of the lock.  This is a bug in
         the guest, and (per POSIX pthread rules) the unlock
         attempt will fail.  So just complain and do nothing
         else. */
      Thread* realOwner = (Thread*)HG_(anyElementOfBag)( lock->heldBy );
      tl_assert(is_sane_Thread(realOwner));
      tl_assert(realOwner != thr);
      tl_assert(!HG_(elemWS)( univ_lsets, thr->locksetA, (Word)lock ));
      tl_assert(!HG_(elemWS)( univ_lsets, thr->locksetW, (Word)lock ));
      record_error_UnlockForeign( thr, realOwner, lock );
      goto error;
   }

   /* Ok, we hold the lock 'n' times. */
   tl_assert(n >= 1);

   lockN_release( lock, thr );

   n--;
   tl_assert(n >= 0);

   if (n > 0) {
      tl_assert(lock->heldBy);
      tl_assert(n == HG_(elemBag)( lock->heldBy, (Word)thr )); 
      /* We still hold the lock.  So either it's a recursive lock 
         or a rwlock which is currently r-held. */
      tl_assert(lock->kind == LK_mbRec
                || (lock->kind == LK_rdwr && !lock->heldW));
      tl_assert(HG_(elemWS)( univ_lsets, thr->locksetA, (Word)lock ));
      if (lock->heldW)
         tl_assert(HG_(elemWS)( univ_lsets, thr->locksetW, (Word)lock ));
      else
         tl_assert(!HG_(elemWS)( univ_lsets, thr->locksetW, (Word)lock ));
   } else {
      /* We no longer hold the lock. */
      if (lock->heldBy) {
         tl_assert(0 == HG_(elemBag)( lock->heldBy, (Word)thr ));
      }
      /* update this thread's lockset accordingly. */
      thr->locksetA
         = HG_(delFromWS)( univ_lsets, thr->locksetA, (Word)lock );
      thr->locksetW
         = HG_(delFromWS)( univ_lsets, thr->locksetW, (Word)lock );
   }
   /* fall through */

  error:
   tl_assert(is_sane_LockN(lock));
}


/*--------- Event handlers proper (evh__* functions) ---------*/

/* What is the Thread* for the currently running thread?  This is
   absolutely performance critical.  We receive notifications from the
   core for client code starts/stops, and cache the looked-up result
   in 'current_Thread'.  Hence, for the vast majority of requests,
   finding the current thread reduces to a read of a global variable,
   provided get_current_Thread_in_C_C is inlined.

   Outside of client code, current_Thread is NULL, and presumably
   any uses of it will cause a segfault.  Hence:

   - for uses definitely within client code, use
     get_current_Thread_in_C_C.

   - for all other uses, use get_current_Thread.
*/

static Thread* current_Thread = NULL;

static void evh__start_client_code ( ThreadId tid, ULong nDisp ) {
   if (0) VG_(printf)("start %d %llu\n", (Int)tid, nDisp);
   tl_assert(current_Thread == NULL);
   current_Thread = map_threads_lookup( tid );
   tl_assert(current_Thread != NULL);
}
static void evh__stop_client_code ( ThreadId tid, ULong nDisp ) {
   if (0) VG_(printf)(" stop %d %llu\n", (Int)tid, nDisp);
   tl_assert(current_Thread != NULL);
   current_Thread = NULL;
}
static inline Thread* get_current_Thread_in_C_C ( void ) {
   return current_Thread;
}
static inline Thread* get_current_Thread ( void ) {
   ThreadId coretid;
   Thread*  thr;
   thr = get_current_Thread_in_C_C();
   if (LIKELY(thr))
      return thr;
   /* evidently not in client code.  Do it the slow way. */
   coretid = VG_(get_running_tid)();
   /* FIXME: get rid of the following kludge.  It exists because
      evim__new_mem is called during initialisation (as notification
      of initial memory layout) and VG_(get_running_tid)() returns
      VG_INVALID_THREADID at that point. */
   if (coretid == VG_INVALID_THREADID)
      coretid = 1; /* KLUDGE */
   thr = map_threads_lookup( coretid );
   return thr;
}

static
void evh__new_mem ( Addr a, SizeT len ) {
   if (SHOW_EVENTS >= 2)
      VG_(printf)("evh__new_mem(%p, %lu)\n", (void*)a, len );
   shadow_mem_make_New( get_current_Thread(), a, len );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__new_mem-post");
}

static
void evh__new_mem_w_perms ( Addr a, SizeT len, 
                            Bool rr, Bool ww, Bool xx ) {
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__new_mem_w_perms(%p, %lu, %d,%d,%d)\n",
                  (void*)a, len, (Int)rr, (Int)ww, (Int)xx );
   if (rr || ww || xx)
      shadow_mem_make_New( get_current_Thread(), a, len );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__new_mem_w_perms-post");
}

static
void evh__set_perms ( Addr a, SizeT len,
                      Bool rr, Bool ww, Bool xx ) {
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__set_perms(%p, %lu, %d,%d,%d)\n",
                  (void*)a, len, (Int)rr, (Int)ww, (Int)xx );
   /* Hmm.  What should we do here, that actually makes any sense?
      Let's say: if neither readable nor writable, then declare it
      NoAccess, else leave it alone. */
   if (!(rr || ww))
      shadow_mem_make_NoAccess( get_current_Thread(), a, len );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__set_perms-post");
}

static
void evh__die_mem ( Addr a, SizeT len ) {
   if (SHOW_EVENTS >= 2)
      VG_(printf)("evh__die_mem(%p, %lu)\n", (void*)a, len );
   shadow_mem_make_NoAccess( get_current_Thread(), a, len );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__die_mem-post");
}

static
void evh__pre_thread_ll_create ( ThreadId parent, ThreadId child )
{
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__pre_thread_ll_create(p=%d, c=%d)\n",
                  (Int)parent, (Int)child );

   if (parent != VG_INVALID_THREADID) {
      Thread*   thr_p;
      Thread*   thr_c;
      SegmentID segid_c;
      Segment*  seg_c;

      tl_assert(is_sane_ThreadId(parent));
      tl_assert(is_sane_ThreadId(child));
      tl_assert(parent != child);

      thr_p = map_threads_maybe_lookup( parent );
      thr_c = map_threads_maybe_lookup( child );

      tl_assert(thr_p != NULL);
      tl_assert(thr_c == NULL);

      /* Create a new thread record for the child. */
      // FIXME: code duplication from init_data_structures
      segid_c = alloc_SegmentID();
      seg_c   = mk_Segment( NULL/*thr*/, NULL/*prev*/, NULL/*other*/ );
      map_segments_add( segid_c, seg_c );

      /* a Thread for the new thread ... */
      thr_c = mk_Thread( segid_c );
      seg_c->thr = thr_c;

      /* and bind it in the thread-map table */
      map_threads[child] = thr_c;

      /* Record where the parent is so we can later refer to this in
         error messages.

         On amd64-linux, this entails a nasty glibc-2.5 specific hack.
         The stack snapshot is taken immediately after the parent has
         returned from its sys_clone call.  Unfortunately there is no
         unwind info for the insn following "syscall" - reading the
         glibc sources confirms this.  So we ask for a snapshot to be
         taken as if RIP was 3 bytes earlier, in a place where there
         is unwind info.  Sigh.
      */
      { Word first_ip_delta = 0;
#       if defined(VGP_amd64_linux)
        first_ip_delta = -3;
#       endif
        thr_c->created_at = VG_(record_ExeContext)(parent, first_ip_delta);
      }

      /* Now, mess with segments. */ 
      if (clo_happens_before >= 1) {
         /* Make the child's new segment depend on the parent */
         seg_c->other = map_segments_lookup( thr_p->csegid );
         seg_c->other_hint = 'c';
         seg_c->vts = tick_VTS( thr_c, seg_c->other->vts );
         tl_assert(seg_c->prev == NULL);
         /* and start a new segment for the parent. */
         { SegmentID new_segid = 0; /* bogus */
           Segment*  new_seg   = NULL;
           evhH__start_new_segment_for_thread( &new_segid, &new_seg, 
                                               thr_p );
           tl_assert(is_sane_SegmentID(new_segid));
           tl_assert(is_sane_Segment(new_seg));
           new_seg->vts = tick_VTS( thr_p, new_seg->prev->vts );
           tl_assert(new_seg->other == NULL);
         }
      }
   }

   if (clo_sanity_flags & SCE_THREADS)
      all__sanity_check("evh__pre_thread_create-post");
}

static
void evh__pre_thread_ll_exit ( ThreadId quit_tid )
{
   Int     nHeld;
   Thread* thr_q;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__pre_thread_ll_exit(thr=%d)\n",
                  (Int)quit_tid );

   /* quit_tid has disappeared without joining to any other thread.
      Therefore there is no synchronisation event associated with its
      exit and so we have to pretty much treat it as if it was still
      alive but mysteriously making no progress.  That is because, if
      we don't know when it really exited, then we can never say there
      is a point in time when we're sure the thread really has
      finished, and so we need to consider the possibility that it
      lingers indefinitely and continues to interact with other
      threads. */
   /* However, it might have rendezvous'd with a thread that called
      pthread_join with this one as arg, prior to this point (that's
      how NPTL works).  In which case there has already been a prior
      sync event.  So in any case, just let the thread exit.  On NPTL,
      all thread exits go through here. */
   tl_assert(is_sane_ThreadId(quit_tid));
   thr_q = map_threads_maybe_lookup( quit_tid );
   tl_assert(thr_q != NULL);

   /* Complain if this thread holds any locks. */
   nHeld = HG_(cardinalityWS)( univ_lsets, thr_q->locksetA );
   tl_assert(nHeld >= 0);
   if (nHeld > 0) {
      HChar buf[80];
      VG_(sprintf)(buf, "Exiting thread still holds %d lock%s",
                        nHeld, nHeld > 1 ? "s" : "");
      record_error_Misc( thr_q, buf );
   }

   /* About the only thing we do need to do is clear the map_threads
      entry, in order that the Valgrind core can re-use it. */
   map_threads_delete( quit_tid );

   if (clo_sanity_flags & SCE_THREADS)
      all__sanity_check("evh__pre_thread_ll_exit-post");
}

static
void evh__HG_PTHREAD_JOIN_POST ( ThreadId stay_tid, Thread* quit_thr )
{
   Int      stats_SMs, stats_SMs_scanned, stats_reExcls;
   Addr     ga;
   SecMap*  sm;
   Thread*  thr_s;
   Thread*  thr_q;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__post_thread_join(stayer=%d, quitter=%p)\n",
                  (Int)stay_tid, quit_thr );

   tl_assert(is_sane_ThreadId(stay_tid));

   thr_s = map_threads_maybe_lookup( stay_tid );
   thr_q = quit_thr;
   tl_assert(thr_s != NULL);
   tl_assert(thr_q != NULL);
   tl_assert(thr_s != thr_q);

   if (clo_happens_before >= 1) {
      /* Start a new segment for the stayer */
      SegmentID new_segid = 0; /* bogus */
      Segment*  new_seg   = NULL;
      evhH__start_new_segment_for_thread( &new_segid, &new_seg, thr_s );
      tl_assert(is_sane_SegmentID(new_segid));
      tl_assert(is_sane_Segment(new_seg));
      /* and make it depend on the quitter's last segment */
      tl_assert(new_seg->other == NULL);
      new_seg->other = map_segments_lookup( thr_q->csegid );
      new_seg->other_hint = 'j';
      tl_assert(new_seg->thr == thr_s);
      new_seg->vts = tickL_and_joinR_VTS( thr_s, new_seg->prev->vts,
                                                 new_seg->other->vts );
   }

   // FIXME: error-if: exiting thread holds any locks
   //        or should evh__pre_thread_ll_exit do that?

   /* Delete thread from ShM/ShR thread sets and restore Excl states
      where appropriate */

   /* When Thread(t) joins to Thread(u):

      scan all shadow memory.  For each ShM/ShR thread set, replace
      't' in each set with 'u'.  If this results in a singleton 'u',
      change the state to Excl(u->csegid).

      Optimisation: tag each SecMap with a superset of the union of
      the thread sets in the SecMap.  Then if the tag set does not
      include 't' then the SecMap can be skipped, because there is no
      't' to change to anything else.

      Problem is that the tag set needs to be updated often, after
      every ShR/ShM store.  (that increases the thread set of the
      shadow value.)

      --> Compromise.  Tag each SecMap with a .mbHasShared bit which
          must be set true if any ShR/ShM on the page.  Set this for
          any transitions into ShR/ShM on the page.  Then skip page if
          not set.

      .mbHasShared bits are (effectively) cached in cache_shmem.
      Hence that must be flushed before we can safely consult them.

      Since we're modifying the backing store, we also need to
      invalidate cache_shmem, so that subsequent memory references get
      up to date shadow values.
   */
   shmem__flush_and_invalidate_scache();

   stats_SMs = stats_SMs_scanned = stats_reExcls = 0;
   HG_(initIterFM)( map_shmem );
   while (HG_(nextIterFM)( map_shmem,
                           (Word*)(void*)&ga, (Word*)(void*)&sm )) {
      SecMapIter itr;
      UInt*      w32p;
      tl_assert(sm);
      stats_SMs++;
      /* Skip this SecMap if the summary bit indicates it is safe to
         do so. */
      if (!sm->mbHasShared)
         continue;
      stats_SMs_scanned++;
      initSecMapIter( &itr );
      while (stepSecMapIter( &w32p, &itr, sm )) {
         Bool isM;
         UInt wnew, wold, lset_old, tset_old, tset_new;
         wold = *w32p;
         if (!is_SHVAL_Sh(wold))
            continue;
         isM = is_SHVAL_ShM(wold);
         lset_old = un_SHVAL_Sh_lset(wold);
         tset_old = un_SHVAL_Sh_tset(wold);
         /* Subst thr_q -> thr_s in the thread set.  Longwindedly, if
            thr_q is in the set, delete it and add thr_s; else leave
            it alone.  FIXME: is inefficient - make a special
            substInWS method for this. */
         tset_new 
            = HG_(elemWS)( univ_tsets, tset_old, (Word)thr_q )
              ? HG_(addToWS)(
                   univ_tsets, 
                   HG_(delFromWS)( univ_tsets, tset_old, (Word)thr_q ),
                   (Word)thr_s 
                )
              : tset_old;

         tl_assert(HG_(cardinalityWS)(univ_tsets, tset_new) 
                   <= HG_(cardinalityWS)(univ_tsets, tset_old));

         if (0) {
            VG_(printf)("smga %p: old 0x%x new 0x%x   ",
                        ga, tset_old, tset_new);
            HG_(ppWS)( univ_tsets, tset_old );
            VG_(printf)("  -->  ");
            HG_(ppWS)( univ_tsets, tset_new );
            VG_(printf)("\n");
         }
         if (HG_(isSingletonWS)( univ_tsets, tset_new, (Word)thr_s )) {
            /* This word returns to Excl state */
            wnew = mk_SHVAL_Excl(thr_s->csegid);
            stats_reExcls++;
         } else {
            wnew = isM ? mk_SHVAL_ShM(tset_new, lset_old)
                       : mk_SHVAL_ShR(tset_new, lset_old);
         }
         *w32p = wnew;
      }
   }
   HG_(doneIterFM)( map_shmem );

   if (SHOW_EXPENSIVE_STUFF)
      VG_(printf)("evh__post_thread_join: %d SMs, "
                  "%d scanned, %d re-Excls\n", 
                  stats_SMs, stats_SMs_scanned, stats_reExcls);

   /* This holds because, at least when using NPTL as the thread
      library, we should be notified the low level thread exit before
      we hear of any join event on it.  The low level exit
      notification feeds through into evh__pre_thread_ll_exit,
      which should clear the map_threads entry for it.  Hence we
      expect there to be no map_threads entry at this point. */
   tl_assert( map_threads_maybe_reverse_lookup_SLOW(thr_q)
              == VG_INVALID_THREADID);

   if (clo_sanity_flags & SCE_THREADS)
      all__sanity_check("evh__post_thread_join-post");
}

static
void evh__pre_mem_read ( CorePart part, ThreadId tid, Char* s, 
                         Addr a, SizeT size) {
   if (SHOW_EVENTS >= 2
       || (SHOW_EVENTS >= 1 && size != 1))
      VG_(printf)("evh__pre_mem_read(ctid=%d, \"%s\", %p, %lu)\n", 
                  (Int)tid, s, (void*)a, size );
   shadow_mem_read_range( map_threads_lookup(tid), a, size);
   if (size >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__pre_mem_read-post");
}

static
void evh__pre_mem_read_asciiz ( CorePart part, ThreadId tid,
                                Char* s, Addr a ) {
   Int len;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__pre_mem_asciiz(ctid=%d, \"%s\", %p)\n", 
                  (Int)tid, s, (void*)a );
   // FIXME: think of a less ugly hack
   len = VG_(strlen)( (Char*) a );
   shadow_mem_read_range( map_threads_lookup(tid), a, len+1 );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__pre_mem_read_asciiz-post");
}

static
void evh__pre_mem_write ( CorePart part, ThreadId tid, Char* s,
                          Addr a, SizeT size ) {
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__pre_mem_write(ctid=%d, \"%s\", %p, %lu)\n", 
                  (Int)tid, s, (void*)a, size );
   shadow_mem_write_range( map_threads_lookup(tid), a, size);
   if (size >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__pre_mem_write-post");
}

static
void evh__new_mem_heap ( Addr a, SizeT len, Bool is_inited ) {
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__new_mem_heap(%p, %lu, inited=%d)\n", 
                  (void*)a, len, (Int)is_inited );
   // FIXME: this is kinda stupid
   if (is_inited) {
      shadow_mem_make_New(get_current_Thread(), a, len);
   } else {
      shadow_mem_make_New(get_current_Thread(), a, len);
   }
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__pre_mem_read-post");
}

static
void evh__die_mem_heap ( Addr a, SizeT len ) {
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__die_mem_heap(%p, %lu)\n", (void*)a, len );
   shadow_mem_make_NoAccess( get_current_Thread(), a, len );
   if (len >= SCE_BIGRANGE_T && (clo_sanity_flags & SCE_BIGRANGE))
      all__sanity_check("evh__pre_mem_read-post");
}

// thread async exit?

static VG_REGPARM(1)
void evh__mem_help_read_1(Addr a) {
   shadow_mem_read8( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_read_2(Addr a) {
   shadow_mem_read16( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_read_4(Addr a) {
   shadow_mem_read32( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_read_8(Addr a) {
   shadow_mem_read64( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(2)
void evh__mem_help_read_N(Addr a, SizeT size) {
   shadow_mem_read_range( get_current_Thread_in_C_C(), a, size );
}

static VG_REGPARM(1)
void evh__mem_help_write_1(Addr a) {
   shadow_mem_write8( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_write_2(Addr a) {
   shadow_mem_write16( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_write_4(Addr a) {
   shadow_mem_write32( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(1)
void evh__mem_help_write_8(Addr a) {
   shadow_mem_write64( get_current_Thread_in_C_C(), a, 0/*unused*/ );
}
static VG_REGPARM(2)
void evh__mem_help_write_N(Addr a, SizeT size) {
   shadow_mem_write_range( get_current_Thread_in_C_C(), a, size );
}

static void evh__bus_lock(void) {
   Thread* thr;
   if (0) VG_(printf)("evh__bus_lock()\n");
   thr = get_current_Thread();
   tl_assert(thr); /* cannot fail - Thread* must already exist */
   evhH__post_thread_w_acquires_lock( thr, LK_nonRec, (Addr)&__bus_lock );
}
static void evh__bus_unlock(void) {
   Thread* thr;
   if (0) VG_(printf)("evh__bus_unlock()\n");
   thr = get_current_Thread();
   tl_assert(thr); /* cannot fail - Thread* must already exist */
   evhH__pre_thread_releases_lock( thr, (Addr)&__bus_lock, False/*!isRDWR*/ );
}


/* -------------- events to do with mutexes -------------- */

/* EXPOSITION only: by intercepting lock init events we can show the
   user where the lock was initialised, rather than only being able to
   show where it was first locked.  Intercepting lock initialisations
   is not necessary for the basic operation of the race checker. */
static
void evh__HG_PTHREAD_MUTEX_INIT_POST( ThreadId tid, 
                                      void* mutex, Word mbRec )
{
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_MUTEX_INIT_POST(ctid=%d, mbRec=%ld, %p)\n", 
                  (Int)tid, mbRec, (void*)mutex );
   tl_assert(mbRec == 0 || mbRec == 1);
   map_locks_lookup_or_create( mbRec ? LK_mbRec : LK_nonRec,
                               (Addr)mutex, tid );
   if (clo_sanity_flags & SCE_LOCKS)
      all__sanity_check("evh__hg_PTHREAD_MUTEX_INIT_POST");
}

static
void evh__HG_PTHREAD_MUTEX_DESTROY_PRE( ThreadId tid, void* mutex )
{
   Thread* thr;
   Lock*   lk;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_MUTEX_DESTROY_PRE(ctid=%d, %p)\n", 
                  (Int)tid, (void*)mutex );

   thr = map_threads_maybe_lookup( tid );
   /* cannot fail - Thread* must already exist */
   tl_assert( is_sane_Thread(thr) );

   lk = map_locks_maybe_lookup( (Addr)mutex );

   if (lk == NULL || (lk->kind != LK_nonRec && lk->kind != LK_mbRec)) {
      record_error_Misc( thr,
                         "pthread_mutex_destroy with invalid argument" );
   }

   if (lk) {
      tl_assert( is_sane_LockN(lk) );
      tl_assert( lk->guestaddr == (Addr)mutex );
      if (lk->heldBy) {
         /* Basically act like we unlocked the lock */
         record_error_Misc( thr, "pthread_mutex_destroy of a locked mutex" );
         /* remove lock from locksets of all owning threads */
         remove_Lock_from_locksets_of_all_owning_Threads( lk );
         HG_(deleteBag)( lk->heldBy );
         lk->heldBy = NULL;
         lk->heldW = False;
         lk->acquired_at = NULL;
      }
      tl_assert( !lk->heldBy );
      tl_assert( is_sane_LockN(lk) );
   }

   if (clo_sanity_flags & SCE_LOCKS)
      all__sanity_check("evh__hg_PTHREAD_MUTEX_DESTROY_PRE");
}

static void evh__HG_PTHREAD_MUTEX_LOCK_PRE ( ThreadId tid,
                                             void* mutex, Word isTryLock )
{
   /* Just check the mutex is sane; nothing else to do. */
   // 'mutex' may be invalid - not checked by wrapper
   Thread* thr;
   Lock*   lk;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_MUTEX_LOCK_PRE(ctid=%d, mutex=%p)\n", 
                  (Int)tid, (void*)mutex );

   tl_assert(isTryLock == 0 || isTryLock == 1);
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   lk = map_locks_maybe_lookup( (Addr)mutex );

   if (lk && (lk->kind == LK_rdwr)) {
      record_error_Misc( thr, "pthread_mutex_lock with a "
                              "pthread_rwlock_t* argument " );
   }

   if ( lk 
        && isTryLock == 0
        && (lk->kind == LK_nonRec || lk->kind == LK_rdwr)
        && lk->heldBy
        && lk->heldW
        && HG_(elemBag)( lk->heldBy, (Word)thr ) > 0 ) {
      /* uh, it's a non-recursive lock and we already w-hold it, and
         this is a real lock operation (not a speculative "tryLock"
         kind of thing).  Duh.  Deadlock coming up; but at least
         produce an error message. */
      record_error_Misc( thr, "Attempt to re-lock a "
                              "non-recursive lock I already hold" );
   }
}

static void evh__HG_PTHREAD_MUTEX_LOCK_POST ( ThreadId tid, void* mutex )
{
   // only called if the real library call succeeded - so mutex is sane
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_PTHREAD_MUTEX_LOCK_POST(ctid=%d, mutex=%p)\n", 
                  (Int)tid, (void*)mutex );

   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   evhH__post_thread_w_acquires_lock( 
      thr, 
      LK_mbRec, /* if not known, create new lock with this LockKind */
      (Addr)mutex
   );
}

static void evh__HG_PTHREAD_MUTEX_UNLOCK_PRE ( ThreadId tid, void* mutex )
{
   // 'mutex' may be invalid - not checked by wrapper
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_PTHREAD_MUTEX_UNLOCK_PRE(ctid=%d, mutex=%p)\n", 
                  (Int)tid, (void*)mutex );

   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   evhH__pre_thread_releases_lock( thr, (Addr)mutex, False/*!isRDWR*/ );
}

static void evh__HG_PTHREAD_MUTEX_UNLOCK_POST ( ThreadId tid, void* mutex )
{
   // only called if the real library call succeeded - so mutex is sane
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_MUTEX_UNLOCK_POST(ctid=%d, mutex=%p)\n", 
                  (Int)tid, (void*)mutex );
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // anything we should do here?
}


/* --------------- events to do with CVs --------------- */

/* A mapping from CV to the thread segment which has most recently
   signalled/broadcasted on it.  This makes it possible to create
   thread segments to model happens-before events arising from CV
   signallings/broadcasts.
*/

/* pthread_mutex_cond* -> Segment* */
static WordFM* map_cond_to_Segment = NULL;

static void map_cond_to_Segment_INIT ( void ) {
   if (UNLIKELY(map_cond_to_Segment == NULL)) {
      map_cond_to_Segment = HG_(newFM)( hg_zalloc, hg_free, NULL );
      tl_assert(map_cond_to_Segment != NULL);
   }
}

static void evh__HG_PTHREAD_COND_SIGNAL_PRE ( ThreadId tid, void* cond )
{
   /* 'tid' has signalled on 'cond'.  Start a new segment for this
      thread, and make a binding from 'cond' to our old segment in the
      mapping.  This is later used by other thread(s) which
      successfully exit from a pthread_cond_wait on the same cv; then
      they know what the signalling segment was, so a dependency edge
      back to it can be constructed. */

   Thread*   thr;
   SegmentID new_segid;
   Segment*  new_seg;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_PTHREAD_COND_SIGNAL_PRE(ctid=%d, cond=%p)\n", 
                  (Int)tid, (void*)cond );

   map_cond_to_Segment_INIT();
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // error-if: mutex is bogus
   // error-if: mutex is not locked

   if (clo_happens_before >= 2) {
      /* create a new segment ... */
      new_segid = 0; /* bogus */
      new_seg   = NULL;
      evhH__start_new_segment_for_thread( &new_segid, &new_seg, thr );
      tl_assert( is_sane_SegmentID(new_segid) );
      tl_assert( is_sane_Segment(new_seg) );
      tl_assert( new_seg->thr == thr );
      tl_assert( is_sane_Segment(new_seg->prev) );
      tl_assert( new_seg->prev->vts );
      new_seg->vts = tick_VTS( new_seg->thr, new_seg->prev->vts );

      /* ... and add the binding. */
      HG_(addToFM)( map_cond_to_Segment, (Word)cond,
                                         (Word)(new_seg->prev) );
   }
}

/* returns True if it reckons 'mutex' is valid and held by this
   thread, else False */
static Bool evh__HG_PTHREAD_COND_WAIT_PRE ( ThreadId tid,
                                            void* cond, void* mutex )
{
   Thread* thr;
   Lock*   lk;
   Bool    lk_valid = True;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_COND_WAIT_PRE"
                  "(ctid=%d, cond=%p, mutex=%p)\n", 
                  (Int)tid, (void*)cond, (void*)mutex );

   map_cond_to_Segment_INIT();
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   lk = map_locks_maybe_lookup( (Addr)mutex );

   /* Check for stupid mutex arguments.  There are various ways to be
      a bozo.  Only complain once, though, even if more than one thing
      is wrong. */
   if (lk == NULL) {
      lk_valid = False;
      record_error_Misc( 
         thr, 
         "pthread_cond_{timed}wait called with invalid mutex" );
   } else {
      tl_assert( is_sane_LockN(lk) );
      if (lk->kind == LK_rdwr) {
         lk_valid = False;
         record_error_Misc( 
            thr, "pthread_cond_{timed}wait called with mutex "
                 "of type pthread_rwlock_t*" );
      } else
         if (lk->heldBy == NULL) {
         lk_valid = False;
         record_error_Misc( 
            thr, "pthread_cond_{timed}wait called with un-held mutex");
      } else
      if (lk->heldBy != NULL
          && HG_(elemBag)( lk->heldBy, (Word)thr ) == 0) {
         lk_valid = False;
         record_error_Misc( 
            thr, "pthread_cond_{timed}wait called with mutex "
                 "held by a different thread" );
      }
   }

   // error-if: cond is also associated with a different mutex

   return lk_valid;
}

static void evh__HG_PTHREAD_COND_WAIT_POST ( ThreadId tid,
                                             void* cond, void* mutex )
{
   /* A pthread_cond_wait(cond, mutex) completed successfully.  Start
      a new segment for this thread.  Look up the signalling-segment
      for the 'cond' in the mapping, and add a dependency edge from
      the new segment back to it. */

   Thread*   thr;
   SegmentID new_segid;
   Segment*  new_seg;
   Segment*  signalling_seg;
   Bool      found;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_PTHREAD_COND_WAIT_POST"
                  "(ctid=%d, cond=%p, mutex=%p)\n", 
                  (Int)tid, (void*)cond, (void*)mutex );

   map_cond_to_Segment_INIT();
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // error-if: cond is also associated with a different mutex

   if (clo_happens_before >= 2) {
      /* create a new segment ... */
      new_segid = 0; /* bogus */
      new_seg   = NULL;
      evhH__start_new_segment_for_thread( &new_segid, &new_seg, thr );
      tl_assert( is_sane_SegmentID(new_segid) );
      tl_assert( is_sane_Segment(new_seg) );
      tl_assert( new_seg->thr == thr );
      tl_assert( is_sane_Segment(new_seg->prev) );
      tl_assert( new_seg->other == NULL);

      /* and find out which thread signalled us; then add a dependency
         edge back to it. */
      signalling_seg = NULL;
      found = HG_(lookupFM)( map_cond_to_Segment, 
                             NULL, (Word*)(void*)&signalling_seg,
                                   (Word)cond );
      if (found) {
         tl_assert(is_sane_Segment(signalling_seg));
         tl_assert(new_seg->prev);
         tl_assert(new_seg->prev->vts);
         new_seg->other      = signalling_seg;
         new_seg->other_hint = 's';
         tl_assert(new_seg->other->vts);
         new_seg->vts = tickL_and_joinR_VTS( 
                           new_seg->thr, 
                           new_seg->prev->vts,
                           new_seg->other->vts );
      } else {
         /* Hmm.  How can a wait on 'cond' succeed if nobody signalled
            it?  If this happened it would surely be a bug in the
            threads library.  Or one of those fabled "spurious
            wakeups". */
         record_error_Misc( thr, "Bug in libpthread: pthread_cond_wait "
                                 "succeeded on"
                                 " without prior pthread_cond_post");
         tl_assert(new_seg->prev->vts);
         new_seg->vts = tick_VTS( new_seg->thr, new_seg->prev->vts );
      }
   }
}


/* -------------- events to do with rwlocks -------------- */

/* EXPOSITION only */
static
void evh__HG_PTHREAD_RWLOCK_INIT_POST( ThreadId tid, void* rwl )
{
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_RWLOCK_INIT_POST(ctid=%d, %p)\n", 
                  (Int)tid, (void*)rwl );
   map_locks_lookup_or_create( LK_rdwr, (Addr)rwl, tid );
   if (clo_sanity_flags & SCE_LOCKS)
      all__sanity_check("evh__hg_PTHREAD_RWLOCK_INIT_POST");
}

static
void evh__HG_PTHREAD_RWLOCK_DESTROY_PRE( ThreadId tid, void* rwl )
{
   Thread* thr;
   Lock*   lk;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_RWLOCK_DESTROY_PRE(ctid=%d, %p)\n", 
                  (Int)tid, (void*)rwl );

   thr = map_threads_maybe_lookup( tid );
   /* cannot fail - Thread* must already exist */
   tl_assert( is_sane_Thread(thr) );

   lk = map_locks_maybe_lookup( (Addr)rwl );

   if (lk == NULL || lk->kind != LK_rdwr) {
      record_error_Misc( thr,
                         "pthread_rwlock_destroy with invalid argument" );
   }

   if (lk) {
      tl_assert( is_sane_LockN(lk) );
      tl_assert( lk->guestaddr == (Addr)rwl );
      if (lk->heldBy) {
         /* Basically act like we unlocked the lock */
         record_error_Misc( thr, "pthread_rwlock_destroy of a locked mutex" );
         /* remove lock from locksets of all owning threads */
         remove_Lock_from_locksets_of_all_owning_Threads( lk );
         HG_(deleteBag)( lk->heldBy );
         lk->heldBy = NULL;
         lk->heldW = False;
      }
      tl_assert( !lk->heldBy );
      tl_assert( is_sane_LockN(lk) );
   }

   if (clo_sanity_flags & SCE_LOCKS)
      all__sanity_check("evh__hg_PTHREAD_RWLOCK_DESTROY_PRE");
}

static 
void evh__HG_PTHREAD_RWLOCK_LOCK_PRE ( ThreadId tid, void* rwl, Word isW )
{
   /* Just check the rwl is sane; nothing else to do. */
   // 'rwl' may be invalid - not checked by wrapper
   Thread* thr;
   Lock*   lk;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_RWLOCK_LOCK_PRE(ctid=%d, isW=%d, %p)\n", 
                  (Int)tid, (Int)isW, (void*)rwl );

   tl_assert(isW == 0 || isW == 1); /* assured us by wrapper */
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   lk = map_locks_maybe_lookup( (Addr)rwl );
   if ( lk 
        && (lk->kind == LK_nonRec || lk->kind == LK_mbRec) ) {
      /* Wrong kind of lock.  Duh.  */
      record_error_Misc( thr, "pthread_rwlock_{rd,rw}lock with a "
                              "pthread_mutex_t* argument " );
   }
}

static 
void evh__HG_PTHREAD_RWLOCK_LOCK_POST ( ThreadId tid, void* rwl, Word isW )
{
   // only called if the real library call succeeded - so mutex is sane
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_RWLOCK_LOCK_POST(ctid=%d, isW=%d, %p)\n", 
                  (Int)tid, (Int)isW, (void*)rwl );

   tl_assert(isW == 0 || isW == 1); /* assured us by wrapper */
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   (isW ? evhH__post_thread_w_acquires_lock 
        : evhH__post_thread_r_acquires_lock)( 
      thr, 
      LK_rdwr, /* if not known, create new lock with this LockKind */
      (Addr)rwl
   );
}

static void evh__HG_PTHREAD_RWLOCK_UNLOCK_PRE ( ThreadId tid, void* rwl )
{
   // 'rwl' may be invalid - not checked by wrapper
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_PTHREAD_RWLOCK_UNLOCK_PRE(ctid=%d, rwl=%p)\n", 
                  (Int)tid, (void*)rwl );

   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   evhH__pre_thread_releases_lock( thr, (Addr)rwl, True/*isRDWR*/ );
}

static void evh__HG_PTHREAD_RWLOCK_UNLOCK_POST ( ThreadId tid, void* rwl )
{
   // only called if the real library call succeeded - so mutex is sane
   Thread* thr;
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__hg_PTHREAD_RWLOCK_UNLOCK_POST(ctid=%d, rwl=%p)\n", 
                  (Int)tid, (void*)rwl );
   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // anything we should do here?
}


/* --------------- events to do with semaphores --------------- */

/* This is similar but not identical the handling for condition
   variables. */

/* For each semaphore, we maintain a stack of Segments.  When a 'post'
   operation is done on a semaphore (unlocking, essentially), a new
   segment is created for the posting thread, and the old segment is
   pushed on the semaphore's stack.

   Later, when a (probably different) thread completes 'wait' on the
   semaphore, we pop a Segment off the semaphore's stack (which should
   be nonempty).  We start a new segment for the thread and make it
   also depend on the just-popped segment.  This mechanism creates
   dependencies between posters and waiters of the semaphore.

   It may not be necessary to use a stack - perhaps a bag of Segments
   would do.  But we do need to keep track of how many unused-up posts
   have happened for the semaphore.

   Imagine T1 and T2 both post once on a semphore S, and T3 waits
   twice on S.  T3 cannot complete its waits without both T1 and T2
   posting.  The above mechanism will ensure that T3 acquires
   dependencies on both T1 and T2.
*/

/* sem_t* -> XArray* Segment* */
static WordFM* map_sem_to_Segment_stack = NULL;

static void map_sem_to_Segment_stack_INIT ( void ) {
   if (map_sem_to_Segment_stack == NULL) {
      map_sem_to_Segment_stack = HG_(newFM)( hg_zalloc, hg_free, NULL );
      tl_assert(map_sem_to_Segment_stack != NULL);
   }
}

static void push_Segment_for_sem ( void* sem, Segment* seg ) {
   XArray* xa;
   tl_assert(seg);
   map_sem_to_Segment_stack_INIT();
   if (HG_(lookupFM)( map_sem_to_Segment_stack, 
                      NULL, (Word*)(void*)&xa, (Word)sem )) {
      tl_assert(xa);
      VG_(addToXA)( xa, &seg );
   } else {
      xa = VG_(newXA)( hg_zalloc, hg_free, sizeof(Segment*) );
      VG_(addToXA)( xa, &seg );
      HG_(addToFM)( map_sem_to_Segment_stack, (Word)sem, (Word)xa );
   }
}

static Segment* mb_pop_Segment_for_sem ( void* sem ) {
   XArray*  xa;
   Segment* seg;
   map_sem_to_Segment_stack_INIT();
   if (HG_(lookupFM)( map_sem_to_Segment_stack, 
                      NULL, (Word*)(void*)&xa, (Word)sem )) {
      /* xa is the stack for this semaphore. */
      Word sz = VG_(sizeXA)( xa );
      tl_assert(sz >= 0);
      if (sz == 0)
         return NULL; /* odd, the stack is empty */
      seg = *(Segment**)VG_(indexXA)( xa, sz-1 );
      tl_assert(seg);
      VG_(dropTailXA)( xa, 1 );
      return seg;
   } else {
      /* hmm, that's odd.  No stack for this semaphore. */
      return NULL;
   }
}

static void evh__HG_POSIX_SEM_ZAPSTACK ( ThreadId tid, void* sem )
{
   Segment* seg;

   /* Empty out the semaphore's segment stack.  Occurs at
      sem_init and sem_destroy time. */
   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_POSIX_SEM_ZAPSTACK(ctid=%d, sem=%p)\n", 
                  (Int)tid, (void*)sem );

   /* This is stupid, but at least it's easy. */
   do {
     seg = mb_pop_Segment_for_sem( sem );
   } while (seg);

   tl_assert(!seg);
}

static void evh__HG_POSIX_SEMPOST_PRE ( ThreadId tid, void* sem )
{
   /* 'tid' has posted on 'sem'.  Start a new segment for this thread,
      and push the old segment on a stack of segments associated with
      'sem'.  This is later used by other thread(s) which successfully
      exit from a sem_wait on the same sem; then they know what the
      posting segment was, so a dependency edge back to it can be
      constructed. */

   Thread*   thr;
   SegmentID new_segid;
   Segment*  new_seg;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_POSIX_SEMPOST_PRE(ctid=%d, sem=%p)\n", 
                  (Int)tid, (void*)sem );

   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // error-if: sem is bogus

   if (clo_happens_before >= 2) {
      /* create a new segment ... */
      new_segid = 0; /* bogus */
      new_seg   = NULL;
      evhH__start_new_segment_for_thread( &new_segid, &new_seg, thr );
      tl_assert( is_sane_SegmentID(new_segid) );
      tl_assert( is_sane_Segment(new_seg) );
      tl_assert( new_seg->thr == thr );
      tl_assert( is_sane_Segment(new_seg->prev) );
      tl_assert( new_seg->prev->vts );
      new_seg->vts = tick_VTS( new_seg->thr, new_seg->prev->vts );

      /* ... and add the binding. */
      push_Segment_for_sem( sem, new_seg->prev );
   }
}

static void evh__HG_POSIX_SEMWAIT_POST ( ThreadId tid, void* sem )
{
   /* A sem_wait(sem) completed successfully.  Start a new segment for
      this thread.  Pop the posting-segment for the 'sem' in the
      mapping, and add a dependency edge from the new segment back to
      it. */

   Thread*   thr;
   SegmentID new_segid;
   Segment*  new_seg;
   Segment*  posting_seg;

   if (SHOW_EVENTS >= 1)
      VG_(printf)("evh__HG_POSIX_SEMWAIT_POST(ctid=%d, sem=%p)\n", 
                  (Int)tid, (void*)sem );

   thr = map_threads_maybe_lookup( tid );
   tl_assert(thr); /* cannot fail - Thread* must already exist */

   // error-if: sem is bogus

   if (clo_happens_before >= 2) {
      /* create a new segment ... */
      new_segid = 0; /* bogus */
      new_seg   = NULL;
      evhH__start_new_segment_for_thread( &new_segid, &new_seg, thr );
      tl_assert( is_sane_SegmentID(new_segid) );
      tl_assert( is_sane_Segment(new_seg) );
      tl_assert( new_seg->thr == thr );
      tl_assert( is_sane_Segment(new_seg->prev) );
      tl_assert( new_seg->other == NULL);

      /* and find out which thread posted last on sem; then add a
         dependency edge back to it. */
      posting_seg = mb_pop_Segment_for_sem( sem );
      if (posting_seg) {
         tl_assert(is_sane_Segment(posting_seg));
         tl_assert(new_seg->prev);
         tl_assert(new_seg->prev->vts);
         new_seg->other      = posting_seg;
         new_seg->other_hint = 'S';
         tl_assert(new_seg->other->vts);
         new_seg->vts = tickL_and_joinR_VTS( 
                           new_seg->thr, 
                           new_seg->prev->vts,
                           new_seg->other->vts );
      } else {
         /* Hmm.  How can a wait on 'sem' succeed if nobody posted to
            it?  If this happened it would surely be a bug in the
            threads library. */
         record_error_Misc( thr, "Bug in libpthread: sem_wait succeeded on"
                                 " semaphore without prior sem_post");
         tl_assert(new_seg->prev->vts);
         new_seg->vts = tick_VTS( new_seg->thr, new_seg->prev->vts );
      }
   }
}


/*--------------------------------------------------------------*/
/*--- Lock acquisition order monitoring                      ---*/
/*--------------------------------------------------------------*/

/* FIXME: here are some optimisations still to do in
          laog__pre_thread_acquires_lock.

   The graph is structured so that if L1 --*--> L2 then L1 must be
   acquired before L2.

   The common case is that some thread T holds (eg) L1 L2 and L3 and
   is repeatedly acquiring and releasing Ln, and there is no ordering
   error in what it is doing.  Hence it repeatly:

   (1) searches laog to see if Ln --*--> {L1,L2,L3}, which always 
       produces the answer No (because there is no error).

   (2) adds edges {L1,L2,L3} --> Ln to laog, which are already present
       (because they already got added the first time T acquired Ln).

   Hence cache these two events:

   (1) Cache result of the query from last time.  Invalidate the cache
       any time any edges are added to or deleted from laog.

   (2) Cache these add-edge requests and ignore them if said edges
       have already been added to laog.  Invalidate the cache any time
       any edges are deleted from laog.
*/

typedef
   struct {
      WordSetID inns; /* in univ_laog */
      WordSetID outs; /* in univ_laog */
   }
   LAOGLinks;

/* lock order acquisition graph */
static WordFM* laog = NULL; /* WordFM Lock* LAOGLinks* */

/* EXPOSITION ONLY: for each edge in 'laog', record the two places
   where that edge was created, so that we can show the user later if
   we need to. */
typedef
   struct {
      Addr        src_ga; /* Lock guest addresses for */
      Addr        dst_ga; /* src/dst of the edge */
      ExeContext* src_ec; /* And corresponding places where that */
      ExeContext* dst_ec; /* ordering was established */
   }
   LAOGLinkExposition;

static Word cmp_LAOGLinkExposition ( Word llx1W, Word llx2W ) {
   /* Compare LAOGLinkExposition*s by (src_ga,dst_ga) field pair. */
   LAOGLinkExposition* llx1 = (LAOGLinkExposition*)llx1W;
   LAOGLinkExposition* llx2 = (LAOGLinkExposition*)llx2W;
   if (llx1->src_ga < llx2->src_ga) return -1;
   if (llx1->src_ga > llx2->src_ga) return  1;
   if (llx1->dst_ga < llx2->dst_ga) return -1;
   if (llx1->dst_ga > llx2->dst_ga) return  1;
   return 0;
}

static WordFM* laog_exposition = NULL; /* WordFM LAOGLinkExposition* NULL */
/* end EXPOSITION ONLY */


static void laog__show ( Char* who ) {
   Word i, ws_size;
   Word* ws_words;
   Lock* me;
   LAOGLinks* links;
   VG_(printf)("laog (requested by %s) {\n", who);
   HG_(initIterFM)( laog );
   me = NULL;
   links = NULL;
   while (HG_(nextIterFM)( laog, (Word*)(void*)&me,
                                 (Word*)(void*)&links )) {
      tl_assert(me);
      tl_assert(links);
      VG_(printf)("   node %p:\n", me);
      HG_(getPayloadWS)( &ws_words, &ws_size, univ_laog, links->inns );
      for (i = 0; i < ws_size; i++)
         VG_(printf)("      inn %p\n", ws_words[i] );
      HG_(getPayloadWS)( &ws_words, &ws_size, univ_laog, links->outs );
      for (i = 0; i < ws_size; i++)
         VG_(printf)("      out %p\n", ws_words[i] );
      me = NULL;
      links = NULL;
   }
   HG_(doneIterFM)( laog );
   VG_(printf)("}\n");
}

__attribute__((noinline))
static void laog__add_edge ( Lock* src, Lock* dst ) {
   Word       keyW;
   LAOGLinks* links;
   Bool       presentF, presentR;
   if (0) VG_(printf)("laog__add_edge %p %p\n", src, dst);

   /* Take the opportunity to sanity check the graph.  Record in
      presentF if there is already a src->dst mapping in this node's
      forwards links, and presentR if there is already a src->dst
      mapping in this node's backwards links.  They should agree!
      Also, we need to know whether the edge was already present so as
      to decide whether or not to update the link details mapping.  We
      can compute presentF and presentR essentially for free, so may
      as well do this always. */
   presentF = presentR = False;

   /* Update the out edges for src */
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)src )) {
      WordSetID outs_new;
      tl_assert(links);
      tl_assert(keyW == (Word)src);
      outs_new = HG_(addToWS)( univ_laog, links->outs, (Word)dst );
      presentF = outs_new == links->outs;
      links->outs = outs_new;
   } else {
      links = hg_zalloc(sizeof(LAOGLinks));
      links->inns = HG_(emptyWS)( univ_laog );
      links->outs = HG_(singletonWS)( univ_laog, (Word)dst );
      HG_(addToFM)( laog, (Word)src, (Word)links );
   }
   /* Update the in edges for dst */
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)dst )) {
      WordSetID inns_new;
      tl_assert(links);
      tl_assert(keyW == (Word)dst);
      inns_new = HG_(addToWS)( univ_laog, links->inns, (Word)src );
      presentR = inns_new == links->inns;
      links->inns = inns_new;
   } else {
      links = hg_zalloc(sizeof(LAOGLinks));
      links->inns = HG_(singletonWS)( univ_laog, (Word)src );
      links->outs = HG_(emptyWS)( univ_laog );
      HG_(addToFM)( laog, (Word)dst, (Word)links );
   }

   tl_assert( (presentF && presentR) || (!presentF && !presentR) );

   if (!presentF && src->acquired_at && dst->acquired_at) {
      LAOGLinkExposition expo;
      /* If this edge is entering the graph, and we have acquired_at
         information for both src and dst, record those acquisition
         points.  Hence, if there is later a violation of this
         ordering, we can show the user the two places in which the
         required src-dst ordering was previously established. */
      if (0) VG_(printf)("acquire edge %p %p\n", 
                         src->guestaddr, dst->guestaddr);
      expo.src_ga = src->guestaddr;
      expo.dst_ga = dst->guestaddr;
      expo.src_ec = NULL;
      expo.dst_ec = NULL;
      tl_assert(laog_exposition);
      if (HG_(lookupFM)( laog_exposition, NULL, NULL, (Word)&expo )) {
         /* we already have it; do nothing */
      } else {
         LAOGLinkExposition* expo2 = hg_zalloc(sizeof(LAOGLinkExposition));
         expo2->src_ga = src->guestaddr;
         expo2->dst_ga = dst->guestaddr;
         expo2->src_ec = src->acquired_at;
         expo2->dst_ec = dst->acquired_at;
         HG_(addToFM)( laog_exposition, (Word)expo2, (Word)NULL );
      }
   }
}

__attribute__((noinline))
static void laog__del_edge ( Lock* src, Lock* dst ) {
   Word       keyW;
   LAOGLinks* links;
   if (0) VG_(printf)("laog__del_edge %p %p\n", src, dst);
   /* Update the out edges for src */
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)src )) {
      tl_assert(links);
      tl_assert(keyW == (Word)src);
      links->outs = HG_(delFromWS)( univ_laog, links->outs, (Word)dst );
   }
   /* Update the in edges for dst */
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)dst )) {
      tl_assert(links);
      tl_assert(keyW == (Word)dst);
      links->inns = HG_(delFromWS)( univ_laog, links->inns, (Word)src );
   }
}

__attribute__((noinline))
static WordSetID /* in univ_laog */ laog__succs ( Lock* lk ) {
   Word       keyW;
   LAOGLinks* links;
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)lk )) {
      tl_assert(links);
      tl_assert(keyW == (Word)lk);
      return links->outs;
   } else {
      return HG_(emptyWS)( univ_laog );
   }
}

__attribute__((noinline))
static WordSetID /* in univ_laog */ laog__preds ( Lock* lk ) {
   Word       keyW;
   LAOGLinks* links;
   keyW  = 0;
   links = NULL;
   if (HG_(lookupFM)( laog, &keyW, (Word*)(void*)&links, (Word)lk )) {
      tl_assert(links);
      tl_assert(keyW == (Word)lk);
      return links->inns;
   } else {
      return HG_(emptyWS)( univ_laog );
   }
}

__attribute__((noinline))
static void laog__sanity_check ( Char* who ) {
   Word i, ws_size;
   Word* ws_words;
   Lock* me;
   LAOGLinks* links;
   if ( !laog )
      return; /* nothing much we can do */
   HG_(initIterFM)( laog );
   me = NULL;
   links = NULL;
   if (0) VG_(printf)("laog sanity check\n");
   while (HG_(nextIterFM)( laog, (Word*)(void*)&me,
                                 (Word*)(void*)&links )) {
      tl_assert(me);
      tl_assert(links);
      HG_(getPayloadWS)( &ws_words, &ws_size, univ_laog, links->inns );
      for (i = 0; i < ws_size; i++) {
         if ( ! HG_(elemWS)( univ_laog, 
                             laog__succs( (Lock*)ws_words[i] ), 
                             (Word)me ))
            goto bad;
      }
      HG_(getPayloadWS)( &ws_words, &ws_size, univ_laog, links->outs );
      for (i = 0; i < ws_size; i++) {
         if ( ! HG_(elemWS)( univ_laog, 
                             laog__preds( (Lock*)ws_words[i] ), 
                             (Word)me ))
            goto bad;
      }
      me = NULL;
      links = NULL;
   }
   HG_(doneIterFM)( laog );
   return;

  bad:
   VG_(printf)("laog__sanity_check(%s) FAILED\n", who);
   laog__show(who);
   tl_assert(0);
}

/* If there is a path in laog from 'src' to any of the elements in
   'dst', return an arbitrarily chosen element of 'dst' reachable from
   'src'.  If no path exist from 'src' to any element in 'dst', return
   NULL. */
__attribute__((noinline))
static
Lock* laog__do_dfs_from_to ( Lock* src, WordSetID dsts /* univ_lsets */ )
{
   Lock*     ret;
   Word      i, ssz;
   XArray*   stack;   /* of Lock* */
   WordFM*   visited; /* Lock* -> void, iow, Set(Lock*) */
   Lock*     here;
   WordSetID succs;
   Word      succs_size;
   Word*     succs_words;
   //laog__sanity_check();

   /* If the destination set is empty, we can never get there from
      'src' :-), so don't bother to try */
   if (HG_(isEmptyWS)( univ_lsets, dsts ))
      return NULL;

   ret     = NULL;
   stack   = VG_(newXA)( hg_zalloc, hg_free, sizeof(Lock*) );
   visited = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxedcmp*/ );

   (void) VG_(addToXA)( stack, &src );

   while (True) {

      ssz = VG_(sizeXA)( stack );

      if (ssz == 0) { ret = NULL; break; }

      here = *(Lock**) VG_(indexXA)( stack, ssz-1 );
      VG_(dropTailXA)( stack, 1 );

      if (HG_(elemWS)( univ_lsets, dsts, (Word)here )) { ret = here; break; }

      if (HG_(lookupFM)( visited, NULL, NULL, (Word)here ))
         continue;

      HG_(addToFM)( visited, (Word)here, 0 );

      succs = laog__succs( here );
      HG_(getPayloadWS)( &succs_words, &succs_size, univ_laog, succs );
      for (i = 0; i < succs_size; i++)
         (void) VG_(addToXA)( stack, &succs_words[i] );
   }

   HG_(deleteFM)( visited, NULL, NULL );
   VG_(deleteXA)( stack );
   return ret;
}


/* Thread 'thr' is acquiring 'lk'.  Check for inconsistent ordering
   between 'lk' and the locks already held by 'thr' and issue a
   complaint if so.  Also, update the ordering graph appropriately.
*/
__attribute__((noinline))
static void laog__pre_thread_acquires_lock ( 
               Thread* thr, /* NB: BEFORE lock is added */
               Lock*   lk
            )
{
   Word*    ls_words;
   Word     ls_size, i;
   Lock*    other;

   /* It may be that 'thr' already holds 'lk' and is recursively
      relocking in.  In this case we just ignore the call. */
   /* NB: univ_lsets really is correct here */
   if (HG_(elemWS)( univ_lsets, thr->locksetA, (Word)lk ))
      return;

   if (!laog)
      laog = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxedcmp*/ );
   if (!laog_exposition)
      laog_exposition = HG_(newFM)( hg_zalloc, hg_free, 
                                    cmp_LAOGLinkExposition );

   /* First, the check.  Complain if there is any path in laog from lk
      to any of the locks already held by thr, since if any such path
      existed, it would mean that previously lk was acquired before
      (rather than after, as we are doing here) at least one of those
      locks.
   */
   other = laog__do_dfs_from_to(lk, thr->locksetA);
   if (other) {
      LAOGLinkExposition key, *found;
      /* So we managed to find a path lk --*--> other in the graph,
         which implies that 'lk' should have been acquired before
         'other' but is in fact being acquired afterwards.  We present
         the lk/other arguments to record_error_LockOrder in the order
         in which they should have been acquired. */
      /* Go look in the laog_exposition mapping, to find the allocation
         points for this edge, so we can show the user. */
      key.src_ga = lk->guestaddr;
      key.dst_ga = other->guestaddr;
      key.src_ec = NULL;
      key.dst_ec = NULL;
      found = NULL;
      if (HG_(lookupFM)( laog_exposition,
                         (Word*)(void*)&found, NULL, (Word)&key )) {
         tl_assert(found != &key);
         tl_assert(found->src_ga == key.src_ga);
         tl_assert(found->dst_ga == key.dst_ga);
         tl_assert(found->src_ec);
         tl_assert(found->dst_ec);
         record_error_LockOrder( thr, 
                                 lk->guestaddr, other->guestaddr,
                                 found->src_ec, found->dst_ec );
      } else {
         /* Hmm.  This can't happen (can it?) */
         record_error_LockOrder( thr, 
                                 lk->guestaddr,        other->guestaddr,
                                 NULL, NULL );
      }
   }

   /* Second, add to laog the pairs
        (old, lk)  |  old <- locks already held by thr
      Since both old and lk are currently held by thr, their acquired_at
      fields must be non-NULL.
   */
   tl_assert(lk->acquired_at);
   HG_(getPayloadWS)( &ls_words, &ls_size, univ_lsets, thr->locksetA );
   for (i = 0; i < ls_size; i++) {
      Lock* old = (Lock*)ls_words[i];
      tl_assert(old->acquired_at);
      laog__add_edge( old, lk );
   }

   /* Why "except_Locks" ?  We're here because a lock is being
      acquired by a thread, and we're in an inconsistent state here.
      See the call points in evhH__post_thread_{r,w}_acquires_lock.
      When called in this inconsistent state, locks__sanity_check duly
      barfs. */
   if (clo_sanity_flags & SCE_LAOG)
      all_except_Locks__sanity_check("laog__pre_thread_acquires_lock-post");
}


/* Delete from 'laog' any pair mentioning a lock in locksToDelete */

__attribute__((noinline))
static void laog__handle_one_lock_deletion ( Lock* lk )
{
   WordSetID preds, succs;
   Word preds_size, succs_size, i, j;
   Word *preds_words, *succs_words;

   preds = laog__preds( lk );
   succs = laog__succs( lk );

   HG_(getPayloadWS)( &preds_words, &preds_size, univ_laog, preds );
   for (i = 0; i < preds_size; i++)
      laog__del_edge( (Lock*)preds_words[i], lk );

   HG_(getPayloadWS)( &succs_words, &succs_size, univ_laog, succs );
   for (j = 0; j < succs_size; j++)
      laog__del_edge( lk, (Lock*)succs_words[j] );

   for (i = 0; i < preds_size; i++) {
      for (j = 0; j < succs_size; j++) {
         if (preds_words[i] != succs_words[j]) {
            /* This can pass unlocked locks to laog__add_edge, since
               we're deleting stuff.  So their acquired_at fields may
               be NULL. */
            laog__add_edge( (Lock*)preds_words[i], (Lock*)succs_words[j] );
         }
      }
   }
}

__attribute__((noinline))
static void laog__handle_lock_deletions (
               WordSetID /* in univ_laog */ locksToDelete
            )
{
   Word  i, ws_size;
   Word* ws_words;

   if (!laog)
      laog = HG_(newFM)( hg_zalloc, hg_free, NULL/*unboxedcmp*/ );
   if (!laog_exposition)
      laog_exposition = HG_(newFM)( hg_zalloc, hg_free, 
                                    cmp_LAOGLinkExposition );

   HG_(getPayloadWS)( &ws_words, &ws_size, univ_lsets, locksToDelete );
   for (i = 0; i < ws_size; i++)
      laog__handle_one_lock_deletion( (Lock*)ws_words[i] );

   if (clo_sanity_flags & SCE_LAOG)
      all__sanity_check("laog__handle_lock_deletions-post");
}


/*--------------------------------------------------------------*/
/*--- Malloc/free replacements                               ---*/
/*--------------------------------------------------------------*/

typedef
   struct {
      void*       next;    /* required by m_hashtable */
      Addr        payload; /* ptr to actual block    */
      SizeT       szB;     /* size requested         */
      ExeContext* where;   /* where it was allocated */
      Thread*     thr;     /* allocating thread      */
   }
   MallocMeta;

/* A hash table of MallocMetas, used to track malloc'd blocks
   (obviously). */
static VgHashTable hg_mallocmeta_table = NULL;


static MallocMeta* new_MallocMeta ( void ) {
   MallocMeta* md = hg_zalloc( sizeof(MallocMeta) );
   tl_assert(md);
   return md;
}
static void delete_MallocMeta ( MallocMeta* md ) {
   hg_free(md);
}


/* Allocate a client block and set up the metadata for it. */

static
void* handle_alloc ( ThreadId tid, 
                     SizeT szB, SizeT alignB, Bool is_zeroed )
{
   Addr        p;
   MallocMeta* md;

   tl_assert( ((SSizeT)szB) >= 0 );
   p = (Addr)VG_(cli_malloc)(alignB, szB);
   if (!p) {
      return NULL;
   }
   if (is_zeroed)
      VG_(memset)((void*)p, 0, szB);

   /* Note that map_threads_lookup must succeed (cannot assert), since
      memory can only be allocated by currently alive threads, hence
      they must have an entry in map_threads. */
   md = new_MallocMeta();
   md->payload = p;
   md->szB     = szB;
   md->where   = VG_(record_ExeContext)( tid, 0 );
   md->thr     = map_threads_lookup( tid );

   VG_(HT_add_node)( hg_mallocmeta_table, (VgHashNode*)md );

   /* Tell the lower level memory wranglers. */
   evh__new_mem_heap( p, szB, is_zeroed );

   return (void*)p;
}

/* Re the checks for less-than-zero (also in hg_cli__realloc below):
   Cast to a signed type to catch any unexpectedly negative args.
   We're assuming here that the size asked for is not greater than
   2^31 bytes (for 32-bit platforms) or 2^63 bytes (for 64-bit
   platforms). */
static void* hg_cli__malloc ( ThreadId tid, SizeT n ) {
   if (((SSizeT)n) < 0) return NULL;
   return handle_alloc ( tid, n, VG_(clo_alignment),
                         /*is_zeroed*/False );
}
static void* hg_cli____builtin_new ( ThreadId tid, SizeT n ) {
   if (((SSizeT)n) < 0) return NULL;
   return handle_alloc ( tid, n, VG_(clo_alignment),
                         /*is_zeroed*/False );
}
static void* hg_cli____builtin_vec_new ( ThreadId tid, SizeT n ) {
   if (((SSizeT)n) < 0) return NULL;
   return handle_alloc ( tid, n, VG_(clo_alignment), 
                         /*is_zeroed*/False );
}
static void* hg_cli__memalign ( ThreadId tid, SizeT align, SizeT n ) {
   if (((SSizeT)n) < 0) return NULL;
   return handle_alloc ( tid, n, align, 
                         /*is_zeroed*/False );
}
static void* hg_cli__calloc ( ThreadId tid, SizeT nmemb, SizeT size1 ) {
   if ( ((SSizeT)nmemb) < 0 || ((SSizeT)size1) < 0 ) return NULL;
   return handle_alloc ( tid, nmemb*size1, VG_(clo_alignment),
                         /*is_zeroed*/True );
}


/* Free a client block, including getting rid of the relevant
   metadata. */

static void handle_free ( ThreadId tid, void* p )
{
   MallocMeta *md, *old_md;
   SizeT      szB;

   /* First see if we can find the metadata for 'p'. */
   md = (MallocMeta*) VG_(HT_lookup)( hg_mallocmeta_table, (UWord)p );
   if (!md)
      return; /* apparently freeing a bogus address.  Oh well. */

   tl_assert(md->payload == (Addr)p);
   szB = md->szB;

   /* Nuke the metadata block */
   old_md = (MallocMeta*)
            VG_(HT_remove)( hg_mallocmeta_table, (UWord)p );
   tl_assert(old_md); /* it must be present - we just found it */
   tl_assert(old_md == md);
   tl_assert(old_md->payload == (Addr)p);

   VG_(cli_free)((void*)old_md->payload);
   delete_MallocMeta(old_md);

   /* Tell the lower level memory wranglers. */
   evh__die_mem_heap( (Addr)p, szB );
}

static void hg_cli__free ( ThreadId tid, void* p ) {
   handle_free(tid, p);
}
static void hg_cli____builtin_delete ( ThreadId tid, void* p ) {
   handle_free(tid, p);
}
static void hg_cli____builtin_vec_delete ( ThreadId tid, void* p ) {
   handle_free(tid, p);
}


static void* hg_cli__realloc ( ThreadId tid, void* payloadV, SizeT new_size )
{
   MallocMeta *md, *md_new, *md_tmp;
   SizeT      i;

   Addr payload = (Addr)payloadV;

   if (((SSizeT)new_size) < 0) return NULL;

   md = (MallocMeta*) VG_(HT_lookup)( hg_mallocmeta_table, (UWord)payload );
   if (!md)
      return NULL; /* apparently realloc-ing a bogus address.  Oh well. */
  
   tl_assert(md->payload == payload);

   if (md->szB == new_size) {
      /* size unchanged */
      md->where = VG_(record_ExeContext)(tid, 0);
      return payloadV;
   }

   if (md->szB > new_size) {
      /* new size is smaller */
      md->szB   = new_size;
      md->where = VG_(record_ExeContext)(tid, 0);
      evh__die_mem_heap( md->payload + new_size, md->szB - new_size );
      return payloadV;
   }

   /* else */ {
      /* new size is bigger */
      Addr p_new = (Addr)VG_(cli_malloc)(VG_(clo_alignment), new_size);

      /* First half kept and copied, second half new */
      // FIXME: shouldn't we use a copier which implements the
      // memory state machine?
      shadow_mem_copy_range( payload, p_new, md->szB );
      evh__new_mem_heap ( p_new + md->szB, new_size - md->szB,
                           /*inited*/False );
      /* FIXME: can anything funny happen here?  specifically, if the
         old range contained a lock, then die_mem_heap will complain.
         Is that the correct behaviour?  Not sure. */
      evh__die_mem_heap( payload, md->szB );

      /* Copy from old to new */
      for (i = 0; i < md->szB; i++)
         ((UChar*)p_new)[i] = ((UChar*)payload)[i];

      /* Because the metadata hash table is index by payload address,
         we have to get rid of the old hash table entry and make a new
         one.  We can't just modify the existing metadata in place,
         because then it would (almost certainly) be in the wrong hash
         chain. */
      md_new = new_MallocMeta();
      *md_new = *md;

      md_tmp = VG_(HT_remove)( hg_mallocmeta_table, payload );
      tl_assert(md_tmp);
      tl_assert(md_tmp == md);

      VG_(cli_free)((void*)md->payload);
      delete_MallocMeta(md);

      /* Update fields */
      md_new->where   = VG_(record_ExeContext)( tid, 0 );
      md_new->szB     = new_size;
      md_new->payload = p_new;
      md_new->thr     = map_threads_lookup( tid );

      /* and add */
      VG_(HT_add_node)( hg_mallocmeta_table, (VgHashNode*)md_new );

      return (void*)p_new;
   }  
}


/*--------------------------------------------------------------*/
/*--- Instrumentation                                        ---*/
/*--------------------------------------------------------------*/

static void instrument_mem_access ( IRSB*   bbOut, 
                                    IRExpr* addr,
                                    Int     szB,
                                    Bool    isStore,
                                    Int     hWordTy_szB )
{
   IRType   tyAddr   = Ity_INVALID;
   HChar*   hName    = NULL;
   void*    hAddr    = NULL;
   Int      regparms = 0;
   IRExpr** argv     = NULL;
   IRDirty* di       = NULL;

   tl_assert(isIRAtom(addr));
   tl_assert(hWordTy_szB == 4 || hWordTy_szB == 8);

   tyAddr = typeOfIRExpr( bbOut->tyenv, addr );
   tl_assert(tyAddr == Ity_I32 || tyAddr == Ity_I64);

   /* So the effective address is in 'addr' now. */
   regparms = 1; // unless stated otherwise
   if (isStore) {
      switch (szB) {
         case 1:
            hName = "evh__mem_help_write_1";
            hAddr = &evh__mem_help_write_1;
            argv = mkIRExprVec_1( addr );
            break;
         case 2:
            hName = "evh__mem_help_write_2";
            hAddr = &evh__mem_help_write_2;
            argv = mkIRExprVec_1( addr );
            break;
         case 4:
            hName = "evh__mem_help_write_4";
            hAddr = &evh__mem_help_write_4;
            argv = mkIRExprVec_1( addr );
            break;
         case 8:
            hName = "evh__mem_help_write_8";
            hAddr = &evh__mem_help_write_8;
            argv = mkIRExprVec_1( addr );
            break;
         default:
            tl_assert(szB > 8 && szB <= 512); /* stay sane */
            regparms = 2;
            hName = "evh__mem_help_write_N";
            hAddr = &evh__mem_help_write_N;
            argv = mkIRExprVec_2( addr, mkIRExpr_HWord( szB ));
            break;
      }
   } else {
      switch (szB) {
         case 1:
            hName = "evh__mem_help_read_1";
            hAddr = &evh__mem_help_read_1;
            argv = mkIRExprVec_1( addr );
            break;
         case 2:
            hName = "evh__mem_help_read_2";
            hAddr = &evh__mem_help_read_2;
            argv = mkIRExprVec_1( addr );
            break;
         case 4:
            hName = "evh__mem_help_read_4";
            hAddr = &evh__mem_help_read_4;
            argv = mkIRExprVec_1( addr );
            break;
         case 8:
            hName = "evh__mem_help_read_8";
            hAddr = &evh__mem_help_read_8;
            argv = mkIRExprVec_1( addr );
            break;
         default: 
            tl_assert(szB > 8 && szB <= 512); /* stay sane */
            regparms = 2;
            hName = "evh__mem_help_read_N";
            hAddr = &evh__mem_help_read_N;
            argv = mkIRExprVec_2( addr, mkIRExpr_HWord( szB ));
            break;
      }
   }

   /* Add the helper. */
   tl_assert(hName);
   tl_assert(hAddr);
   tl_assert(argv);
   di = unsafeIRDirty_0_N( regparms,
                           hName, VG_(fnptr_to_fnentry)( hAddr ),
                           argv );
   addStmtToIRSB( bbOut, IRStmt_Dirty(di) );
}


static void instrument_memory_bus_event ( IRSB* bbOut, IRMBusEvent event )
{
   switch (event) {
      case Imbe_Fence:
         break; /* not interesting */
      case Imbe_BusLock:
      case Imbe_BusUnlock:
         addStmtToIRSB(
            bbOut,
            IRStmt_Dirty(
               unsafeIRDirty_0_N( 
                  0/*regparms*/, 
                  event == Imbe_BusLock ? "evh__bus_lock"
                                        : "evh__bus_unlock",
                  VG_(fnptr_to_fnentry)(
                     event == Imbe_BusLock ? &evh__bus_lock 
                                           : &evh__bus_unlock 
                  ),
                  mkIRExprVec_0() 
               )
            )
         );
         break;
      default:
         tl_assert(0);
   }
}


static
IRSB* hg_instrument ( VgCallbackClosure* closure,
                      IRSB* bbIn,
                      VexGuestLayout* layout,
                      VexGuestExtents* vge,
                      IRType gWordTy, IRType hWordTy )
{
   Int   i;
   IRSB* bbOut;

   if (gWordTy != hWordTy) {
      /* We don't currently support this case. */
      VG_(tool_panic)("host/guest word size mismatch");
   }

   /* Set up BB */
   bbOut           = emptyIRSB();
   bbOut->tyenv    = deepCopyIRTypeEnv(bbIn->tyenv);
   bbOut->next     = deepCopyIRExpr(bbIn->next);
   bbOut->jumpkind = bbIn->jumpkind;

   // Copy verbatim any IR preamble preceding the first IMark
   i = 0;
   while (i < bbIn->stmts_used && bbIn->stmts[i]->tag != Ist_IMark) {
      addStmtToIRSB( bbOut, bbIn->stmts[i] );
      i++;
   }

   for (/*use current i*/; i < bbIn->stmts_used; i++) {
      IRStmt* st = bbIn->stmts[i];
      tl_assert(st);
      tl_assert(isFlatIRStmt(st));
      switch (st->tag) {
         case Ist_NoOp:
         case Ist_AbiHint:
         case Ist_Put:
         case Ist_PutI:
         case Ist_IMark:
         case Ist_Exit:
            /* None of these can contain any memory references. */
            break;

         case Ist_MBE:
            instrument_memory_bus_event( bbOut, st->Ist.MBE.event );
            break;

         case Ist_Store:
            instrument_mem_access( 
               bbOut, 
               st->Ist.Store.addr, 
               sizeofIRType(typeOfIRExpr(bbIn->tyenv, st->Ist.Store.data)),
               True/*isStore*/,
               sizeofIRType(hWordTy)
            );
            break;

         case Ist_WrTmp: {
            IRExpr* data = st->Ist.WrTmp.data;
            if (data->tag == Iex_Load) {
               instrument_mem_access(
                  bbOut,
                  data->Iex.Load.addr,
                  sizeofIRType(data->Iex.Load.ty),
                  False/*!isStore*/,
                  sizeofIRType(hWordTy)
               );
            }
            break;
         }

         case Ist_Dirty: {
            Int      dataSize;
            IRDirty* d = st->Ist.Dirty.details;
            if (d->mFx != Ifx_None) {
               /* This dirty helper accesses memory.  Collect the
                  details. */
               tl_assert(d->mAddr != NULL);
               tl_assert(d->mSize != 0);
               dataSize = d->mSize;
               if (d->mFx == Ifx_Read || d->mFx == Ifx_Modify) {
                  instrument_mem_access( 
                     bbOut, d->mAddr, dataSize, False/*!isStore*/,
                     sizeofIRType(hWordTy)
                  );
               }
               if (d->mFx == Ifx_Write || d->mFx == Ifx_Modify) {
                  instrument_mem_access( 
                     bbOut, d->mAddr, dataSize, True/*isStore*/,
                     sizeofIRType(hWordTy)
                  );
               }
            } else {
               tl_assert(d->mAddr == NULL);
               tl_assert(d->mSize == 0);
            }
            break;
         }

         default:
            tl_assert(0);

      } /* switch (st->tag) */

      addStmtToIRSB( bbOut, st );
   } /* iterate over bbIn->stmts */

   return bbOut;
}


/*----------------------------------------------------------------*/
/*--- Client requests                                          ---*/
/*----------------------------------------------------------------*/

/* Sheesh.  Yet another goddam finite map. */
static WordFM* map_pthread_t_to_Thread = NULL; /* pthread_t -> Thread* */

static void map_pthread_t_to_Thread_INIT ( void ) {
   if (UNLIKELY(map_pthread_t_to_Thread == NULL)) {
      map_pthread_t_to_Thread = HG_(newFM)( hg_zalloc, hg_free, NULL );
      tl_assert(map_pthread_t_to_Thread != NULL);
   }
}


static 
Bool hg_handle_client_request ( ThreadId tid, UWord* args, UWord* ret)
{
   if (!VG_IS_TOOL_USERREQ('H','G',args[0]))
      return False;

   /* Anything that gets past the above check is one of ours, so we
      should be able to handle it. */

   /* default, meaningless return value, unless otherwise set */
   *ret = 0;

   switch (args[0]) {

      /* --- --- User-visible client requests --- --- */

      case VG_USERREQ__HG_CLEAN_MEMORY:
         if (0) VG_(printf)("VG_USERREQ__HG_CLEAN_MEMORY(%p,%d)\n",
                            args[1], args[2]);
         /* Call die_mem to (expensively) tidy up properly, if there
            are any held locks etc in the area */
         if (args[2] > 0) { /* length */
            evh__die_mem(args[1], args[2]);
            /* and then set it to New */
            evh__new_mem(args[1], args[2]);
         }
         break;

      /* --- --- Client requests for Helgrind's use only --- --- */

      /* Some thread is telling us its pthread_t value.  Record the
         binding between that and the associated Thread*, so we can
         later find the Thread* again when notified of a join by the
         thread. */
      case _VG_USERREQ__HG_SET_MY_PTHREAD_T: {
         Thread* my_thr = NULL;
         if (0)
         VG_(printf)("SET_MY_PTHREAD_T (tid %d): pthread_t = %p\n", (Int)tid,
                     (void*)args[1]);
         map_pthread_t_to_Thread_INIT();
         my_thr = map_threads_maybe_lookup( tid );
         /* This assertion should hold because the map_threads (tid to
            Thread*) binding should have been made at the point of
            low-level creation of this thread, which should have
            happened prior to us getting this client request for it.
            That's because this client request is sent from
            client-world from the 'thread_wrapper' function, which
            only runs once the thread has been low-level created. */
         tl_assert(my_thr != NULL);
         /* So now we know that (pthread_t)args[1] is associated with
            (Thread*)my_thr.  Note that down. */
         if (0)
         VG_(printf)("XXXX: bind pthread_t %p to Thread* %p\n",
                     (void*)args[1], (void*)my_thr );
         HG_(addToFM)( map_pthread_t_to_Thread, (Word)args[1], (Word)my_thr );
         break;
      }

      case _VG_USERREQ__HG_PTH_API_ERROR: {
         Thread* my_thr = NULL;
         map_pthread_t_to_Thread_INIT();
         my_thr = map_threads_maybe_lookup( tid );
         tl_assert(my_thr); /* See justification above in SET_MY_PTHREAD_T */
         record_error_PthAPIerror( my_thr, (HChar*)args[1], 
                                           (Word)args[2], (HChar*)args[3] );
         break;
      }

      /* This thread (tid) has completed a join with the quitting
         thread whose pthread_t is in args[1]. */
      case _VG_USERREQ__HG_PTHREAD_JOIN_POST: {
         Thread* thr_q = NULL; /* quitter Thread* */
         Bool    found = False;
         if (0)
         VG_(printf)("NOTIFY_JOIN_COMPLETE (tid %d): quitter = %p\n", (Int)tid,
                     (void*)args[1]);
         map_pthread_t_to_Thread_INIT();
         found = HG_(lookupFM)( map_pthread_t_to_Thread, 
                                NULL, (Word*)(void*)&thr_q, (Word)args[1] );
          /* Can this fail?  It would mean that our pthread_join
             wrapper observed a successful join on args[1] yet that
             thread never existed (or at least, it never lodged an
             entry in the mapping (via SET_MY_PTHREAD_T)).  Which
             sounds like a bug in the threads library. */
         // FIXME: get rid of this assertion; handle properly
         tl_assert(found);
         if (found) {
            if (0)
            VG_(printf)(".................... quitter Thread* = %p\n", 
                        thr_q);
            evh__HG_PTHREAD_JOIN_POST( tid, thr_q );
         }
         break;
      }

      /* EXPOSITION only: by intercepting lock init events we can show
         the user where the lock was initialised, rather than only
         being able to show where it was first locked.  Intercepting
         lock initialisations is not necessary for the basic operation
         of the race checker. */
      case _VG_USERREQ__HG_PTHREAD_MUTEX_INIT_POST:
         evh__HG_PTHREAD_MUTEX_INIT_POST( tid, (void*)args[1], args[2] );
         break;

      case _VG_USERREQ__HG_PTHREAD_MUTEX_DESTROY_PRE:
         evh__HG_PTHREAD_MUTEX_DESTROY_PRE( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_PTHREAD_MUTEX_UNLOCK_PRE:   // pth_mx_t*
         evh__HG_PTHREAD_MUTEX_UNLOCK_PRE( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_PTHREAD_MUTEX_UNLOCK_POST:  // pth_mx_t*
         evh__HG_PTHREAD_MUTEX_UNLOCK_POST( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_PTHREAD_MUTEX_LOCK_PRE:     // pth_mx_t*, Word
         evh__HG_PTHREAD_MUTEX_LOCK_PRE( tid, (void*)args[1], args[2] );
         break;

      case _VG_USERREQ__HG_PTHREAD_MUTEX_LOCK_POST:    // pth_mx_t*
         evh__HG_PTHREAD_MUTEX_LOCK_POST( tid, (void*)args[1] );
         break;

      /* This thread is about to do pthread_cond_signal on the
         pthread_cond_t* in arg[1].  Ditto pthread_cond_broadcast. */
      case _VG_USERREQ__HG_PTHREAD_COND_SIGNAL_PRE:
      case _VG_USERREQ__HG_PTHREAD_COND_BROADCAST_PRE:
         evh__HG_PTHREAD_COND_SIGNAL_PRE( tid, (void*)args[1] );
         break;

      /* Entry into pthread_cond_wait, cond=arg[1], mutex=arg[2].
         Returns a flag indicating whether or not the mutex is believed to be
         valid for this operation. */
      case _VG_USERREQ__HG_PTHREAD_COND_WAIT_PRE: {
         Bool mutex_is_valid
            = evh__HG_PTHREAD_COND_WAIT_PRE( tid, (void*)args[1], 
                                                  (void*)args[2] );
         *ret = mutex_is_valid ? 1 : 0;
         break;
      }

      /* Thread successfully completed pthread_cond_wait, cond=arg[1],
         mutex=arg[2] */
      case _VG_USERREQ__HG_PTHREAD_COND_WAIT_POST:
         evh__HG_PTHREAD_COND_WAIT_POST( tid,
                                         (void*)args[1], (void*)args[2] );
         break;

      case _VG_USERREQ__HG_PTHREAD_RWLOCK_INIT_POST:
         evh__HG_PTHREAD_RWLOCK_INIT_POST( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_PTHREAD_RWLOCK_DESTROY_PRE:
         evh__HG_PTHREAD_RWLOCK_DESTROY_PRE( tid, (void*)args[1] );
         break;

      /* rwlock=arg[1], isW=arg[2] */
      case _VG_USERREQ__HG_PTHREAD_RWLOCK_LOCK_PRE:
         evh__HG_PTHREAD_RWLOCK_LOCK_PRE( tid, (void*)args[1], args[2] );
         break;

      /* rwlock=arg[1], isW=arg[2] */
      case _VG_USERREQ__HG_PTHREAD_RWLOCK_LOCK_POST:
         evh__HG_PTHREAD_RWLOCK_LOCK_POST( tid, (void*)args[1], args[2] );
         break;

      case _VG_USERREQ__HG_PTHREAD_RWLOCK_UNLOCK_PRE:
         evh__HG_PTHREAD_RWLOCK_UNLOCK_PRE( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_PTHREAD_RWLOCK_UNLOCK_POST:
         evh__HG_PTHREAD_RWLOCK_UNLOCK_POST( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_POSIX_SEMPOST_PRE: /* sem_t* */
         evh__HG_POSIX_SEMPOST_PRE( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_POSIX_SEMWAIT_POST: /* sem_t* */
         evh__HG_POSIX_SEMWAIT_POST( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_POSIX_SEM_ZAPSTACK: /* sem_t* */
         evh__HG_POSIX_SEM_ZAPSTACK( tid, (void*)args[1] );
         break;

      case _VG_USERREQ__HG_GET_MY_SEGMENT: { // -> Segment*
         Thread*   thr;
         SegmentID segid;
         Segment*  seg;
         thr = map_threads_maybe_lookup( tid );
         tl_assert(thr); /* cannot fail */
         segid = thr->csegid;
         tl_assert(is_sane_SegmentID(segid));
         seg = map_segments_lookup( segid );
         tl_assert(seg);
         *ret = (UWord)seg;
         break;
      }

      default:
         /* Unhandled Helgrind client request! */
        tl_assert2(0, "unhandled Helgrind client request!");
   }

   return True;
}


/*----------------------------------------------------------------*/
/*--- Error management                                         ---*/
/*----------------------------------------------------------------*/

/* maps (by value) strings to a copy of them in ARENA_TOOL */
static UWord stats__string_table_queries = 0;
static WordFM* string_table = NULL;
static Word string_table_cmp ( Word s1, Word s2 ) {
   return (Word)VG_(strcmp)( (HChar*)s1, (HChar*)s2 );
}
static HChar* string_table_strdup ( HChar* str ) {
   HChar* copy = NULL;
   stats__string_table_queries++;
   if (!str)
      str = "(null)";
   if (!string_table) {
      string_table = HG_(newFM)( hg_zalloc, hg_free, string_table_cmp );
      tl_assert(string_table);
   }
   if (HG_(lookupFM)( string_table,
                      NULL, (Word*)(void*)&copy, (Word)str )) {
      tl_assert(copy);
      if (0) VG_(printf)("string_table_strdup: %p -> %p\n", str, copy );
      return copy;
   } else {
      copy = VG_(strdup)(str);
      tl_assert(copy);
      HG_(addToFM)( string_table, (Word)copy, (Word)copy );
      return copy;
   }
}

/* maps from Lock .unique fields to LockP*s */
static UWord stats__ga_LockN_to_P_queries = 0;
static WordFM* yaWFM = NULL;
static Word lock_unique_cmp ( Word lk1W, Word lk2W )
{
   Lock* lk1 = (Lock*)lk1W;
   Lock* lk2 = (Lock*)lk2W;
   tl_assert( is_sane_LockNorP(lk1) );
   tl_assert( is_sane_LockNorP(lk2) );
   if (lk1->unique < lk2->unique) return -1;
   if (lk1->unique > lk2->unique) return 1;
   return 0;
}
static Lock* mk_LockP_from_LockN ( Lock* lkn )
{
   Lock* lkp = NULL;
   stats__ga_LockN_to_P_queries++;
   tl_assert( is_sane_LockN(lkn) );
   if (!yaWFM) {
      yaWFM = HG_(newFM)( hg_zalloc, hg_free, lock_unique_cmp );
      tl_assert(yaWFM);
   }
   if (!HG_(lookupFM)( yaWFM, NULL, (Word*)(void*)&lkp, (Word)lkn)) {
      lkp = hg_zalloc( sizeof(Lock) );
      *lkp = *lkn;
      lkp->admin = NULL;
      lkp->magic = LockP_MAGIC;
      /* Forget about the bag of lock holders - don't copy that.
         Also, acquired_at should be NULL whenever heldBy is, and vice
         versa. */
      lkp->heldW  = False;
      lkp->heldBy = NULL;
      lkp->acquired_at = NULL;
      HG_(addToFM)( yaWFM, (Word)lkp, (Word)lkp );
   }
   tl_assert( is_sane_LockP(lkp) );
   return lkp;
}

/* Errors:

      race: program counter
            read or write
            data size
            previous state
            current state

      FIXME: how does state printing interact with lockset gc?
      Are the locksets in prev/curr state always valid?
      Ditto question for the threadsets
          ThreadSets - probably are always valid if Threads
          are never thrown away.
          LockSets - could at least print the lockset elements that
          correspond to actual locks at the time of printing.  Hmm.
*/

/* Error kinds */
typedef
   enum {
      XE_Race=1101,      // race
      XE_FreeMemLock,    // freeing memory containing a locked lock
      XE_UnlockUnlocked, // unlocking a not-locked lock
      XE_UnlockForeign,  // unlocking a lock held by some other thread
      XE_UnlockBogus,    // unlocking an address not known to be a lock
      XE_PthAPIerror,    // error from the POSIX pthreads API
      XE_LockOrder,      // lock order error
      XE_Misc            // misc other error (w/ string to describe it)
   }
   XErrorTag;

/* Extra contexts for kinds */
typedef
   struct  {
      XErrorTag tag;
      union {
         struct {
            Addr  data_addr;
            Int   szB;
            Bool  isWrite;
            UInt  new_state;
            UInt  old_state;
            ExeContext* mb_lastlock;
            Thread* thr;
         } Race;
         struct {
            Thread* thr;  /* doing the freeing */
            Lock*   lock; /* lock which is locked */
         } FreeMemLock;
         struct {
            Thread* thr;  /* doing the unlocking */
            Lock*   lock; /* lock (that is already unlocked) */
         } UnlockUnlocked;
         struct {
            Thread* thr;    /* doing the unlocking */
            Thread* owner;  /* thread that actually holds the lock */
            Lock*   lock;   /* lock (that is held by 'owner') */
         } UnlockForeign;
         struct {
            Thread* thr;     /* doing the unlocking */
            Addr    lock_ga; /* purported address of the lock */
         } UnlockBogus;
         struct {
            Thread* thr; 
            HChar*  fnname; /* persistent, in tool-arena */
            Word    err;    /* pth error code */
            HChar*  errstr; /* persistent, in tool-arena */
         } PthAPIerror;
         struct {
            Thread*     thr;
            Addr        before_ga; /* always locked first in prog. history */
            Addr        after_ga;
            ExeContext* before_ec;
            ExeContext* after_ec;
         } LockOrder;
         struct {
            Thread* thr;
            HChar*  errstr; /* persistent, in tool-arena */
         } Misc;
      } XE;
   }
   XError;

static void init_XError ( XError* xe ) {
   VG_(memset)(xe, 0, sizeof(*xe) );
   xe->tag = XE_Race-1; /* bogus */
}


/* Extensions of suppressions */
typedef
   enum {
      XS_Race=1201, /* race */
      XS_FreeMemLock,
      XS_UnlockUnlocked,
      XS_UnlockForeign,
      XS_UnlockBogus,
      XS_PthAPIerror,
      XS_LockOrder,
      XS_Misc
   }
   XSuppTag;


/* Updates the copy with address info if necessary. */
static UInt hg_update_extra ( Error* err )
{
   XError* extra = (XError*)VG_(get_error_extra)(err);
   tl_assert(extra);
   //if (extra != NULL && Undescribed == extra->addrinfo.akind) {
   //   describe_addr ( VG_(get_error_address)(err), &(extra->addrinfo) );
   //}
   return sizeof(XError);
}

static void record_error_Race ( Thread* thr, 
                                Addr data_addr, Bool isWrite, Int szB,
                                UInt old_sv, UInt new_sv,
                                ExeContext* mb_lastlock ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   init_XError(&xe);
   xe.tag = XE_Race;
   xe.XE.Race.data_addr   = data_addr;
   xe.XE.Race.szB         = szB;
   xe.XE.Race.isWrite     = isWrite;
   xe.XE.Race.new_state   = new_sv;
   xe.XE.Race.old_state   = old_sv;
   xe.XE.Race.mb_lastlock = mb_lastlock;
   xe.XE.Race.thr         = thr;
   // FIXME: tid vs thr
   tl_assert(isWrite == False || isWrite == True);
   tl_assert(szB == 8 || szB == 4 || szB == 2 || szB == 1);
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_Race, data_addr, NULL, &xe );
}

static void record_error_FreeMemLock ( Thread* thr, Lock* lk ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   tl_assert( is_sane_LockN(lk) );
   init_XError(&xe);
   xe.tag = XE_FreeMemLock;
   xe.XE.FreeMemLock.thr  = thr;
   xe.XE.FreeMemLock.lock = mk_LockP_from_LockN(lk);
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_FreeMemLock, 0, NULL, &xe );
}

static void record_error_UnlockUnlocked ( Thread* thr, Lock* lk ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   tl_assert( is_sane_LockN(lk) );
   init_XError(&xe);
   xe.tag = XE_UnlockUnlocked;
   xe.XE.UnlockUnlocked.thr  = thr;
   xe.XE.UnlockUnlocked.lock = mk_LockP_from_LockN(lk);
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_UnlockUnlocked, 0, NULL, &xe );
}

static void record_error_UnlockForeign ( Thread* thr,
                                         Thread* owner, Lock* lk ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   tl_assert( is_sane_Thread(owner) );
   tl_assert( is_sane_LockN(lk) );
   init_XError(&xe);
   xe.tag = XE_UnlockForeign;
   xe.XE.UnlockForeign.thr   = thr;
   xe.XE.UnlockForeign.owner = owner;
   xe.XE.UnlockForeign.lock  = mk_LockP_from_LockN(lk);
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_UnlockForeign, 0, NULL, &xe );
}

static void record_error_UnlockBogus ( Thread* thr, Addr lock_ga ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   init_XError(&xe);
   xe.tag = XE_UnlockBogus;
   xe.XE.UnlockBogus.thr     = thr;
   xe.XE.UnlockBogus.lock_ga = lock_ga;
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_UnlockBogus, 0, NULL, &xe );
}

static 
void record_error_LockOrder ( Thread* thr, Addr before_ga, Addr after_ga,
                              ExeContext* before_ec, ExeContext* after_ec ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   init_XError(&xe);
   xe.tag = XE_LockOrder;
   xe.XE.LockOrder.thr       = thr;
   xe.XE.LockOrder.before_ga = before_ga;
   xe.XE.LockOrder.before_ec = before_ec;
   xe.XE.LockOrder.after_ga  = after_ga;
   xe.XE.LockOrder.after_ec  = after_ec;
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_LockOrder, 0, NULL, &xe );
}

static 
void record_error_PthAPIerror ( Thread* thr, HChar* fnname, 
                                Word err, HChar* errstr ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   tl_assert(fnname);
   tl_assert(errstr);
   init_XError(&xe);
   xe.tag = XE_PthAPIerror;
   xe.XE.PthAPIerror.thr    = thr;
   xe.XE.PthAPIerror.fnname = string_table_strdup(fnname);
   xe.XE.PthAPIerror.err    = err;
   xe.XE.PthAPIerror.errstr = string_table_strdup(errstr);
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_PthAPIerror, 0, NULL, &xe );
}

static void record_error_Misc ( Thread* thr, HChar* errstr ) {
   XError xe;
   tl_assert( is_sane_Thread(thr) );
   tl_assert(errstr);
   init_XError(&xe);
   xe.tag = XE_Misc;
   xe.XE.Misc.thr    = thr;
   xe.XE.Misc.errstr = string_table_strdup(errstr);
   // FIXME: tid vs thr
   VG_(maybe_record_error)( map_threads_reverse_lookup_SLOW(thr),
                            XE_Misc, 0, NULL, &xe );
}

static Bool hg_eq_Error ( VgRes not_used, Error* e1, Error* e2 )
{
   XError *xe1, *xe2;

   tl_assert(VG_(get_error_kind)(e1) == VG_(get_error_kind)(e2));

   xe1 = (XError*)VG_(get_error_extra)(e1);
   xe2 = (XError*)VG_(get_error_extra)(e2);
   tl_assert(xe1);
   tl_assert(xe2);

   switch (VG_(get_error_kind)(e1)) {
      case XE_Race:
         return xe1->XE.Race.szB == xe2->XE.Race.szB
                && xe1->XE.Race.isWrite == xe2->XE.Race.isWrite
                && (clo_cmp_race_err_addrs 
                       ? xe1->XE.Race.data_addr == xe2->XE.Race.data_addr
                       : True);
      case XE_FreeMemLock:
         return xe1->XE.FreeMemLock.thr == xe2->XE.FreeMemLock.thr
                && xe1->XE.FreeMemLock.lock == xe2->XE.FreeMemLock.lock;
      case XE_UnlockUnlocked:
         return xe1->XE.UnlockUnlocked.thr == xe2->XE.UnlockUnlocked.thr
                && xe1->XE.UnlockUnlocked.lock == xe2->XE.UnlockUnlocked.lock;
      case XE_UnlockForeign:
         return xe1->XE.UnlockForeign.thr == xe2->XE.UnlockForeign.thr
                && xe1->XE.UnlockForeign.owner == xe2->XE.UnlockForeign.owner
                && xe1->XE.UnlockForeign.lock == xe2->XE.UnlockForeign.lock;
      case XE_UnlockBogus:
         return xe1->XE.UnlockBogus.thr == xe2->XE.UnlockBogus.thr
                && xe1->XE.UnlockBogus.lock_ga == xe2->XE.UnlockBogus.lock_ga;
      case XE_PthAPIerror:
         return xe1->XE.PthAPIerror.thr == xe2->XE.PthAPIerror.thr
                && 0==VG_(strcmp)(xe1->XE.PthAPIerror.fnname,
                                  xe2->XE.PthAPIerror.fnname)
                && xe1->XE.PthAPIerror.err == xe2->XE.PthAPIerror.err;
      case XE_LockOrder:
         return xe1->XE.LockOrder.thr == xe2->XE.LockOrder.thr;
      case XE_Misc:
         return xe1->XE.Misc.thr == xe2->XE.Misc.thr
                && 0==VG_(strcmp)(xe1->XE.Misc.errstr, xe2->XE.Misc.errstr);
      default:
         tl_assert(0);
   }

   /*NOTREACHED*/
   tl_assert(0);
}

/* Given a WordSetID in univ_tsets (that is, a Thread set ID), produce
   an XArray* with the corresponding Thread*'s sorted by their
   errmsg_index fields.  This is for printing out thread sets in
   repeatable orders, which is important for for repeatable regression
   testing.  The returned XArray* is dynamically allocated (of course)
   and so must be hg_freed by the caller. */
static Int cmp_Thread_by_errmsg_index ( void* thr1V, void* thr2V ) {
   Thread* thr1 = *(Thread**)thr1V;
   Thread* thr2 = *(Thread**)thr2V;
   if (thr1->errmsg_index < thr2->errmsg_index) return -1;
   if (thr1->errmsg_index > thr2->errmsg_index) return  1;
   return 0;
}
static XArray* /* of Thread* */ get_sorted_thread_set ( WordSetID tset )
{
   XArray* xa;
   Word*   ts_words;
   Word    ts_size, i;
   xa = VG_(newXA)( hg_zalloc, hg_free, sizeof(Thread*) );
   tl_assert(xa);
   HG_(getPayloadWS)( &ts_words, &ts_size, univ_tsets, tset );
   tl_assert(ts_words);
   tl_assert(ts_size >= 0);
   /* This isn't a very clever scheme, but we don't expect this to be
      called very often. */
   for (i = 0; i < ts_size; i++) {
      Thread* thr = (Thread*)ts_words[i];
      tl_assert(is_sane_Thread(thr));
      VG_(addToXA)( xa, (void*)&thr );
   }
   tl_assert(ts_size == VG_(sizeXA)( xa ));
   VG_(setCmpFnXA)( xa, cmp_Thread_by_errmsg_index );
   VG_(sortXA)( xa );
   return xa;
}


/* Announce (that is, print the point-of-creation) of the threads in
   'tset'.  Only do this once, as we only want to see these
   announcements once each.  Also, first sort the threads by their
   errmsg_index fields, and show only the first N_THREADS_TO_ANNOUNCE.
   That's because we only want to bother to announce threads
   enumerated by summarise_threadset() below, and that in turn does
   the same: it sorts them and then only shows the first
   N_THREADS_TO_ANNOUNCE. */

static void announce_threadset ( WordSetID tset )
{
   const Word limit = N_THREADS_TO_ANNOUNCE;
   Thread* thr;
   XArray* sorted;
   Word    ts_size, i, loopmax;
   sorted = get_sorted_thread_set( tset );
   ts_size = VG_(sizeXA)( sorted );
   tl_assert(ts_size >= 0);
   loopmax = limit < ts_size  ? limit  : ts_size; /* min(limit, ts_size) */
   tl_assert(loopmax >= 0 && loopmax <= limit);
   for (i = 0; i < loopmax; i++) {
      thr = *(Thread**)VG_(indexXA)( sorted, i );
      tl_assert(is_sane_Thread(thr));
      tl_assert(thr->errmsg_index >= 1);
      if (thr->announced)
         continue;
      if (thr->errmsg_index == 1/*FIXME: this hardwires an assumption
                                  about the identity of the root
                                  thread*/) {
         tl_assert(thr->created_at == NULL);
         VG_(message)(Vg_UserMsg, "Thread #%d is the program's root thread",
                                  thr->errmsg_index);
      } else {
         tl_assert(thr->created_at != NULL);
         VG_(message)(Vg_UserMsg, "Thread #%d was created",
                                  thr->errmsg_index);
         VG_(pp_ExeContext)( thr->created_at );
      }
      VG_(message)(Vg_UserMsg, "");
      thr->announced = True;
   }
   VG_(deleteXA)( sorted );
}
static void announce_one_thread ( Thread* thr ) {
   announce_threadset( HG_(singletonWS)(univ_tsets, (Word)thr ));
}

/* Generate into buf[0 .. nBuf-1] a 1-line summary of a thread set, of
   the form "#1, #3, #77, #78, #79 and 42 others".  The first
   N_THREADS_TO_ANNOUNCE are listed explicitly (as '#n') and the
   leftovers lumped into the 'and n others' bit. */

static void summarise_threadset ( WordSetID tset, Char* buf, UInt nBuf )
{
   const Word limit = N_THREADS_TO_ANNOUNCE;
   Thread* thr;
   XArray* sorted;
   Word    ts_size, i, loopmax;
   UInt    off = 0;
   tl_assert(nBuf > 0);
   tl_assert(nBuf >= 40 + 20*limit);
   tl_assert(buf);
   sorted = get_sorted_thread_set( tset );
   ts_size = VG_(sizeXA)( sorted );
   tl_assert(ts_size >= 0);
   loopmax = limit < ts_size  ? limit  : ts_size; /* min(limit, ts_size) */
   tl_assert(loopmax >= 0 && loopmax <= limit);
   VG_(memset)(buf, 0, nBuf);
   for (i = 0; i < loopmax; i++) {
      thr = *(Thread**)VG_(indexXA)( sorted, i );
      tl_assert(is_sane_Thread(thr));
      tl_assert(thr->errmsg_index >= 1);
      off += VG_(sprintf)(&buf[off], "#%d", (Int)thr->errmsg_index);
      if (i < loopmax-1)
         off += VG_(sprintf)(&buf[off], ", ");
   }
   if (limit < ts_size) {
      Word others = ts_size - limit;
      off += VG_(sprintf)(&buf[off], " and %d other%s", 
                                     (Int)others, others > 1 ? "s" : "");
   }
   tl_assert(off < nBuf);
   tl_assert(buf[nBuf-1] == 0);
   VG_(deleteXA)( sorted );
}

static void hg_pp_Error ( Error* err )
{
   const Bool show_raw_states = False;
   XError *xe = (XError*)VG_(get_error_extra)(err);

   switch (VG_(get_error_kind)(err)) {

   case XE_Misc: {
      tl_assert(xe);
      tl_assert( is_sane_Thread( xe->XE.Misc.thr ) );
      announce_one_thread( xe->XE.Misc.thr );
      VG_(message)(Vg_UserMsg,
                  "Thread #%d: %s",
                  (Int)xe->XE.Misc.thr->errmsg_index,
                  xe->XE.Misc.errstr);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      break;
   }

   case XE_LockOrder: {
      tl_assert(xe);
      tl_assert( is_sane_Thread( xe->XE.LockOrder.thr ) );
      announce_one_thread( xe->XE.LockOrder.thr );
      VG_(message)(Vg_UserMsg,
                  "Thread #%d: lock order \"%p before %p\" violated",
                  (Int)xe->XE.LockOrder.thr->errmsg_index,
                  (void*)xe->XE.LockOrder.before_ga,
                  (void*)xe->XE.LockOrder.after_ga);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      if (xe->XE.LockOrder.before_ec && xe->XE.LockOrder.after_ec) {
         VG_(message)(Vg_UserMsg,
            "  Required order was established by acquisition of lock at %p",
            (void*)xe->XE.LockOrder.before_ga);
         VG_(pp_ExeContext)( xe->XE.LockOrder.before_ec );
         VG_(message)(Vg_UserMsg,
            "  followed by a later acquisition of lock at %p", 
            (void*)xe->XE.LockOrder.after_ga);
         VG_(pp_ExeContext)( xe->XE.LockOrder.after_ec );
      }
      break;
   }

   case XE_PthAPIerror: {
      tl_assert(xe);
      tl_assert( is_sane_Thread( xe->XE.PthAPIerror.thr ) );
      announce_one_thread( xe->XE.PthAPIerror.thr );
      VG_(message)(Vg_UserMsg,
                  "Thread #%d's call to %s failed",
                  (Int)xe->XE.PthAPIerror.thr->errmsg_index,
                  xe->XE.PthAPIerror.fnname);
      VG_(message)(Vg_UserMsg,
                  "   with error code %ld (%s)",
                  xe->XE.PthAPIerror.err,
                  xe->XE.PthAPIerror.errstr);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      break;
   }

   case XE_UnlockBogus: {
      tl_assert(xe);
      tl_assert( is_sane_Thread( xe->XE.UnlockBogus.thr ) );
      announce_one_thread( xe->XE.UnlockBogus.thr );
      VG_(message)(Vg_UserMsg,
                   "Thread #%d unlocked an invalid lock at %p ",
                   (Int)xe->XE.UnlockBogus.thr->errmsg_index,
                   (void*)xe->XE.UnlockBogus.lock_ga);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      break;
   }

   case XE_UnlockForeign: {
      tl_assert(xe);
      tl_assert( is_sane_LockP( xe->XE.UnlockForeign.lock ) );
      tl_assert( is_sane_Thread( xe->XE.UnlockForeign.owner ) );
      tl_assert( is_sane_Thread( xe->XE.UnlockForeign.thr ) );
      announce_one_thread( xe->XE.UnlockForeign.thr );
      announce_one_thread( xe->XE.UnlockForeign.owner );
      VG_(message)(Vg_UserMsg,
                   "Thread #%d unlocked lock at %p "
                   "currently held by thread #%d",
                   (Int)xe->XE.UnlockForeign.thr->errmsg_index,
                   (void*)xe->XE.UnlockForeign.lock->guestaddr,
                   (Int)xe->XE.UnlockForeign.owner->errmsg_index );
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      if (xe->XE.UnlockForeign.lock->appeared_at) {
         VG_(message)(Vg_UserMsg,
                      "  Lock at %p was first observed",
                      (void*)xe->XE.UnlockForeign.lock->guestaddr);
         VG_(pp_ExeContext)( xe->XE.UnlockForeign.lock->appeared_at );
      }
      break;
   }

   case XE_UnlockUnlocked: {
      tl_assert(xe);
      tl_assert( is_sane_LockP( xe->XE.UnlockUnlocked.lock ) );
      tl_assert( is_sane_Thread( xe->XE.UnlockUnlocked.thr ) );
      announce_one_thread( xe->XE.UnlockUnlocked.thr );
      VG_(message)(Vg_UserMsg,
                   "Thread #%d unlocked a not-locked lock at %p ",
                   (Int)xe->XE.UnlockUnlocked.thr->errmsg_index,
                   (void*)xe->XE.UnlockUnlocked.lock->guestaddr);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      if (xe->XE.UnlockUnlocked.lock->appeared_at) {
         VG_(message)(Vg_UserMsg,
                      "  Lock at %p was first observed",
                      (void*)xe->XE.UnlockUnlocked.lock->guestaddr);
         VG_(pp_ExeContext)( xe->XE.UnlockUnlocked.lock->appeared_at );
      }
      break;
   }

   case XE_FreeMemLock: {
      tl_assert(xe);
      tl_assert( is_sane_LockP( xe->XE.FreeMemLock.lock ) );
      tl_assert( is_sane_Thread( xe->XE.FreeMemLock.thr ) );
      announce_one_thread( xe->XE.FreeMemLock.thr );
      VG_(message)(Vg_UserMsg,
                   "Thread #%d deallocated location %p "
                   "containing a locked lock",
                   (Int)xe->XE.FreeMemLock.thr->errmsg_index,
                   (void*)xe->XE.FreeMemLock.lock->guestaddr);
      VG_(pp_ExeContext)( VG_(get_error_where)(err) );
      if (xe->XE.FreeMemLock.lock->appeared_at) {
         VG_(message)(Vg_UserMsg,
                      "  Lock at %p was first observed",
                      (void*)xe->XE.FreeMemLock.lock->guestaddr);
         VG_(pp_ExeContext)( xe->XE.FreeMemLock.lock->appeared_at );
      }
      break;
   }

   case XE_Race: {
      Addr      err_ga;
      Char      old_buf[100], new_buf[100];
      Char      old_tset_buf[140], new_tset_buf[140];
      UInt      old_state, new_state;
      Thread*   thr_acc;
      HChar*    what;
      Int       szB;
      WordSetID tset_to_announce = HG_(emptyWS)( univ_tsets );

      /* First extract some essential info */
      tl_assert(xe);
      old_state = xe->XE.Race.old_state;
      new_state = xe->XE.Race.new_state;
      thr_acc   = xe->XE.Race.thr;
      what      = xe->XE.Race.isWrite ? "write" : "read";
      szB       = xe->XE.Race.szB;
      tl_assert(is_sane_Thread(thr_acc));
      err_ga = VG_(get_error_address)(err);

      /* Format the low level state print descriptions */
      show_shadow_w32(old_buf, sizeof(old_buf), old_state);
      show_shadow_w32(new_buf, sizeof(new_buf), new_state);

      /* Now we have to 'announce' the threadset mentioned in the
         error message, if it hasn't already been announced.
         Unfortunately the precise threadset and error message text
         depends on the nature of the transition involved.  So now
         fall into a case analysis of the error state transitions. */

      /* CASE of Excl -> ShM */
      if (is_SHVAL_Excl(old_state) && is_SHVAL_ShM(new_state)) {
         SegmentID old_segid;
         Segment*  old_seg;
         Thread*   old_thr; 
         WordSetID new_tset;
         old_segid = un_SHVAL_Excl( old_state );
         tl_assert(is_sane_SegmentID(old_segid));
         old_seg = map_segments_lookup( old_segid );
         tl_assert(is_sane_Segment(old_seg));
         tl_assert(old_seg->thr);
         old_thr = old_seg->thr;
         tl_assert(is_sane_Thread(old_thr));

         new_tset = un_SHVAL_ShM_tset(new_state);
         tset_to_announce = HG_(addToWS)( univ_tsets,
                                          new_tset, (Word)old_thr );
         announce_threadset( tset_to_announce );

         VG_(message)(Vg_UserMsg,
                      "Possible data race during %s of size %d at %p",
                      what, szB, err_ga);
         VG_(pp_ExeContext)( VG_(get_error_where)(err) );
         /* pp_AddrInfo(err_addr, &extra->addrinfo); */
         if (show_raw_states)
         VG_(message)(Vg_UserMsg,
                      "  Old state 0x%08x=%s, new state 0x%08x=%s",
                      old_state, old_buf, new_state, new_buf);
         VG_(message)(Vg_UserMsg,
                      "  Old state: owned exclusively by thread #%d",
                      old_thr->errmsg_index);
         // This should always show exactly 2 threads
         summarise_threadset( new_tset, new_tset_buf, sizeof(new_tset_buf) );
         VG_(message)(Vg_UserMsg,
                      "  New state: shared-modified by threads %s",
                      new_tset_buf );
         VG_(message)(Vg_UserMsg,
                      "  Reason:    this thread, #%d, holds no locks at all",
                      thr_acc->errmsg_index);
      }
      else 
      /* Case of ShR/M -> ShM */
      if (is_SHVAL_Sh(old_state) && is_SHVAL_ShM(new_state)) {
         WordSetID old_tset = un_SHVAL_Sh_tset(old_state);
         WordSetID new_tset = un_SHVAL_Sh_tset(new_state);

         tset_to_announce = HG_(unionWS)( univ_tsets, old_tset, new_tset );
         announce_threadset( tset_to_announce );

         VG_(message)(Vg_UserMsg,
                      "Possible data race during %s of size %d at %p",
                      what, szB, err_ga);
         VG_(pp_ExeContext)( VG_(get_error_where)(err) );
         /* pp_AddrInfo(err_addr, &extra->addrinfo); */
         if (show_raw_states)
         VG_(message)(Vg_UserMsg,
                      "  Old state 0x%08x=%s, new state 0x%08x=%s",
                      old_state, old_buf, new_state, new_buf);

         summarise_threadset( old_tset, old_tset_buf, sizeof(old_tset_buf) );
         summarise_threadset( new_tset, new_tset_buf, sizeof(new_tset_buf) );

         VG_(message)(Vg_UserMsg,
                      "  Old state: shared-%s by threads %s", 
                      is_SHVAL_ShM(old_state) ? "modified" : "readonly", 
                      old_tset_buf);
         VG_(message)(Vg_UserMsg,
                      "  New state: shared-modified by threads %s", 
                      new_tset_buf);
         VG_(message)(Vg_UserMsg,
                      "  Reason:    this thread, #%d, holds no "
                      "consistent locks",
                      thr_acc->errmsg_index);
         if (xe->XE.Race.mb_lastlock) {
            VG_(message)(Vg_UserMsg, "  Last consistently used lock for %p was "
                                     "first observed", err_ga);
            VG_(pp_ExeContext)(xe->XE.Race.mb_lastlock);
         } else {
            VG_(message)(Vg_UserMsg, "  Location %p has never been protected "
                                     "by any lock", err_ga);
         }
      }
      /* Hmm, unknown transition.  Just print what we do know. */
      else {
         VG_(message)(Vg_UserMsg,
                      "Possible data race during %s of size %d at %p",
                      what, szB, err_ga);
         VG_(pp_ExeContext)( VG_(get_error_where)(err) );

         //pp_AddrInfo(err_addr, &extra->addrinfo);
         VG_(message)(Vg_UserMsg,
                      "  Old state 0x%08x=%s, new state 0x%08x=%s",
                      old_state, old_buf, new_state, new_buf);
      }

      break; /* case XE_Race */
   } /* case XE_Race */

   default:
      tl_assert(0);
   } /* switch (VG_(get_error_kind)(err)) */
}

static Char* hg_get_error_name ( Error* err )
{
   switch (VG_(get_error_kind)(err)) {
      case XE_Race:           return "Race";
      case XE_FreeMemLock:    return "FreeMemLock";
      case XE_UnlockUnlocked: return "UnlockUnlocked";
      case XE_UnlockForeign:  return "UnlockForeign";
      case XE_UnlockBogus:    return "UnlockBogus";
      case XE_PthAPIerror:    return "PthAPIerror";
      case XE_LockOrder:      return "LockOrder";
      case XE_Misc:           return "Misc";
      default: tl_assert(0); /* fill in missing case */
   }
}

static Bool hg_recognised_suppression ( Char* name, Supp *su )
{
#  define TRY(_name,_xskind)                   \
      if (0 == VG_(strcmp)(name, (_name))) {   \
         VG_(set_supp_kind)(su, (_xskind));    \
         return True;                          \
      }
   TRY("Race",           XS_Race);
   TRY("FreeMemLock",    XS_FreeMemLock);
   TRY("UnlockUnlocked", XS_UnlockUnlocked);
   TRY("UnlockForeign",  XS_UnlockForeign);
   TRY("UnlockBogus",    XS_UnlockBogus);
   TRY("PthAPIerror",    XS_PthAPIerror);
   TRY("LockOrder",      XS_LockOrder);
   TRY("Misc",           XS_Misc);
   return False;
#  undef TRY
}

static Bool hg_read_extra_suppression_info ( Int fd, Char* buf, Int nBuf,
                                             Supp* su )
{
   /* do nothing -- no extra suppression info present.  Return True to
      indicate nothing bad happened. */
   return True;
}

static Bool hg_error_matches_suppression ( Error* err, Supp* su )
{
   switch (VG_(get_supp_kind)(su)) {
   case XS_Race:           return VG_(get_error_kind)(err) == XE_Race;
   case XS_FreeMemLock:    return VG_(get_error_kind)(err) == XE_FreeMemLock;
   case XS_UnlockUnlocked: return VG_(get_error_kind)(err) == XE_UnlockUnlocked;
   case XS_UnlockForeign:  return VG_(get_error_kind)(err) == XE_UnlockForeign;
   case XS_UnlockBogus:    return VG_(get_error_kind)(err) == XE_UnlockBogus;
   case XS_PthAPIerror:    return VG_(get_error_kind)(err) == XE_PthAPIerror;
   case XS_LockOrder:      return VG_(get_error_kind)(err) == XE_LockOrder;
   case XS_Misc:           return VG_(get_error_kind)(err) == XE_Misc;
   //case XS_: return VG_(get_error_kind)(err) == XE_;
   default: tl_assert(0); /* fill in missing cases */
   }
}

static void hg_print_extra_suppression_info ( Error* err )
{
   /* Do nothing */
}


/*----------------------------------------------------------------*/
/*--- Setup                                                    ---*/
/*----------------------------------------------------------------*/

static Bool hg_process_cmd_line_option ( Char* arg )
{
   if      (VG_CLO_STREQ(arg, "--happens-before=none"))
      clo_happens_before = 0;
   else if (VG_CLO_STREQ(arg, "--happens-before=threads"))
      clo_happens_before = 1;
   else if (VG_CLO_STREQ(arg, "--happens-before=all"))
      clo_happens_before = 2;

   else if (VG_CLO_STREQ(arg, "--gen-vcg=no"))
      clo_gen_vcg = 0;
   else if (VG_CLO_STREQ(arg, "--gen-vcg=yes"))
      clo_gen_vcg = 1;
   else if (VG_CLO_STREQ(arg, "--gen-vcg=yes-w-vts"))
      clo_gen_vcg = 2;

   else if (VG_CLO_STREQ(arg, "--cmp-race-err-addrs=no"))
      clo_cmp_race_err_addrs = False;
   else if (VG_CLO_STREQ(arg, "--cmp-race-err-addrs=yes"))
      clo_cmp_race_err_addrs = True;

   else if (VG_CLO_STREQN(13, arg, "--trace-addr=")) {
      clo_trace_addr = VG_(atoll16)(&arg[13]);
      if (clo_trace_level == 0)
         clo_trace_level = 1;
   }
   else VG_BNUM_CLO(arg, "--trace-level", clo_trace_level, 0, 2)

   /* "stuvw" --> stuvw (binary) */
   else if (VG_CLO_STREQN(18, arg, "--tc-sanity-flags=")) {
      Int j;
      char* opt = & arg[18];
   
      if (5 != VG_(strlen)(opt)) {
         VG_(message)(Vg_UserMsg, 
                      "--tc-sanity-flags argument must have 5 digits");
         return False;
      }
      for (j = 0; j < 5; j++) {
         if      ('0' == opt[j]) { /* do nothing */ }
         else if ('1' == opt[j]) clo_sanity_flags |= (1 << (5-1-j));
         else {
            VG_(message)(Vg_UserMsg, "--tc-sanity-flags argument can "
                                     "only contain 0s and 1s");
            return False;
         }
      }
      if (0) VG_(printf)("XXX sanity flags: 0x%x\n", clo_sanity_flags);
   }

   else 
      return VG_(replacement_malloc_process_cmd_line_option)(arg);

   return True;
}

static void hg_print_usage ( void )
{
   VG_(printf)(
"    --happens-before=none|threads|all   [all] consider no events, thread\n"
"      create/join, create/join/cvsignal/cvwait/semwait/post as sync points\n"
"    --trace-addr=0xXXYYZZ     show all state changes for address 0xXXYYZZ\n"
"    --trace-level=0|1|2       verbosity level of --trace-addr [1]\n"
   );
   VG_(replacement_malloc_print_usage)();
}

static void hg_print_debug_usage ( void )
{
   VG_(replacement_malloc_print_debug_usage)();
   VG_(printf)("    --gen-vcg=no|yes|yes-w-vts   show happens-before graph "
               "in .vcg format [no]\n");
   VG_(printf)("    --cmp-race-err-addrs=no|yes  are data addresses in "
               "race errors significant? [no]\n");
   VG_(printf)("    --tc-sanity-flags=<XXXXX> sanity check "
               "  at events (X = 0|1) [00000]\n");
   VG_(printf)("    --tc-sanity-flags values:\n");
   VG_(printf)("       10000   after changes to "
               "lock-order-acquisition-graph\n");
   VG_(printf)("       01000   at memory accesses (NB: not curently used)\n");
   VG_(printf)("       00100   at mem permission setting for "
               "ranges >= %d bytes\n", SCE_BIGRANGE_T);
   VG_(printf)("       00010   at lock/unlock events\n");
   VG_(printf)("       00001   at thread create/join events\n");
}

static void hg_post_clo_init ( void )
{
}

static void hg_fini ( Int exitcode )
{
   if (SHOW_DATA_STRUCTURES)
      pp_everything( PP_ALL, "SK_(fini)" );
   if (clo_sanity_flags)
      all__sanity_check("SK_(fini)");

   if (clo_gen_vcg > 0)
      segments__generate_vcg();

   if (VG_(clo_verbosity) >= 2) {

      if (1) {
         VG_(printf)("\n");
         HG_(ppWSUstats)( univ_tsets, "univ_tsets" );
         VG_(printf)("\n");
         HG_(ppWSUstats)( univ_lsets, "univ_lsets" );
         VG_(printf)("\n");
         HG_(ppWSUstats)( univ_laog,  "univ_laog" );
      }

      VG_(printf)("\n");
      VG_(printf)(" hbefore: %,10lu queries\n",        stats__hbefore_queries);
      VG_(printf)(" hbefore: %,10lu cache 0 hits\n",   stats__hbefore_cache0s);
      VG_(printf)(" hbefore: %,10lu cache > 0 hits\n", stats__hbefore_cacheNs);
      VG_(printf)(" hbefore: %,10lu graph searches\n", stats__hbefore_gsearches);
      VG_(printf)(" hbefore: %,10lu   of which slow\n",
                  stats__hbefore_gsearches - stats__hbefore_gsearchFs);
      VG_(printf)(" hbefore: %,10lu stack high water mark\n",
                  stats__hbefore_stk_hwm);
      VG_(printf)(" hbefore: %,10lu cache invals\n",   stats__hbefore_invals);
      VG_(printf)(" hbefore: %,10lu probes\n",         stats__hbefore_probes);

      VG_(printf)("\n");
      VG_(printf)("        segments: %,8lu Segment objects allocated\n", 
                  stats__mk_Segment);
      VG_(printf)("        locksets: %,8d unique lock sets\n",
                  (Int)HG_(cardinalityWSU)( univ_lsets ));
      VG_(printf)("      threadsets: %,8d unique thread sets\n",
                  (Int)HG_(cardinalityWSU)( univ_tsets ));
      VG_(printf)("       univ_laog: %,8d unique lock sets\n",
                  (Int)HG_(cardinalityWSU)( univ_laog ));

      VG_(printf)("L(ast)L(ock) map: %,8lu inserts (%d map size)\n", 
                  stats__ga_LL_adds,
                  (Int)(ga_to_lastlock ? HG_(sizeFM)( ga_to_lastlock ) : 0) );

      VG_(printf)("  LockN-to-P map: %,8lu queries (%d map size)\n", 
                  stats__ga_LockN_to_P_queries,
                  (Int)(yaWFM ? HG_(sizeFM)( yaWFM ) : 0) );

      VG_(printf)("string table map: %,8lu queries (%d map size)\n", 
                  stats__string_table_queries,
                  (Int)(string_table ? HG_(sizeFM)( string_table ) : 0) );
      VG_(printf)("            LAOG: %,8d map size\n", 
                  (Int)(laog ? HG_(sizeFM)( laog ) : 0));
      VG_(printf)(" LAOG exposition: %,8d map size\n", 
                  (Int)(laog_exposition ? HG_(sizeFM)( laog_exposition ) : 0));
      VG_(printf)("           locks: %,8lu acquires, "
                  "%,lu releases\n",
                  stats__lockN_acquires,
                  stats__lockN_releases
                 );
      VG_(printf)("   sanity checks: %,8lu\n", stats__sanity_checks);

      VG_(printf)("\n");
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_Excl_nochange\n",
                  stats__msm_read_Excl_nochange, stats__msm_write_Excl_nochange);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_Excl_transfer\n",
                  stats__msm_read_Excl_transfer, stats__msm_write_Excl_transfer);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_Excl_to_ShR/ShM\n",
                  stats__msm_read_Excl_to_ShR,   stats__msm_write_Excl_to_ShM);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_ShR_to_ShR/ShM\n",
                  stats__msm_read_ShR_to_ShR,    stats__msm_write_ShR_to_ShM);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_ShM_to_ShM\n",
                  stats__msm_read_ShM_to_ShM,    stats__msm_write_ShM_to_ShM);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_New_to_Excl\n",
                  stats__msm_read_New_to_Excl,   stats__msm_write_New_to_Excl);
      VG_(printf)("     msm: %,12lu %,12lu rd/wr_NoAccess\n",
                  stats__msm_read_NoAccess,      stats__msm_write_NoAccess);

      VG_(printf)("\n");
      VG_(printf)(" secmaps: %,10lu allocd (%,12lu g-a-range)\n",
                  stats__secmaps_allocd,
                  stats__secmap_ga_space_covered);
      VG_(printf)("  linesZ: %,10lu allocd (%,12lu bytes occupied)\n",
                  stats__secmap_linesZ_allocd,
                  stats__secmap_linesZ_bytes);
      VG_(printf)("  linesF: %,10lu allocd (%,12lu bytes occupied)\n",
                  stats__secmap_linesF_allocd,
                  stats__secmap_linesF_bytes);
      VG_(printf)(" secmaps: %,10lu iterator steppings\n",
                  stats__secmap_iterator_steppings);

      VG_(printf)("\n");
      VG_(printf)("   cache: %,lu totrefs (%,lu misses)\n",
                  stats__cache_totrefs, stats__cache_totmisses );
      VG_(printf)("   cache: %,12lu Z-fetch, %,12lu F-fetch\n",
                  stats__cache_Z_fetches, stats__cache_F_fetches );
      VG_(printf)("   cache: %,12lu Z-wback, %,12lu F-wback\n",
                  stats__cache_Z_wbacks, stats__cache_F_wbacks );
      VG_(printf)("   cache: %,12lu invals,  %,12lu flushes\n",
                  stats__cache_invals, stats__cache_flushes );

      VG_(printf)("\n");
      VG_(printf)("   cline: %,10lu normalises\n",
                  stats__cline_normalises );
      VG_(printf)("   cline:  reads 8/4/2/1: %,12lu %,12lu %,12lu %,12lu\n",
                  stats__cline_read64s,
                  stats__cline_read32s,
                  stats__cline_read16s,
                  stats__cline_read8s );
      VG_(printf)("   cline: writes 8/4/2/1: %,12lu %,12lu %,12lu %,12lu\n",
                  stats__cline_write64s,
                  stats__cline_write32s,
                  stats__cline_write16s,
                  stats__cline_write8s );
      VG_(printf)("   cline:   sets 8/4/2/1: %,12lu %,12lu %,12lu %,12lu\n",
                  stats__cline_set64s,
                  stats__cline_set32s,
                  stats__cline_set16s,
                  stats__cline_set8s );
      VG_(printf)("   cline: get1s %,lu, copy1s %,lu\n",
                  stats__cline_get8s, stats__cline_copy8s );
      VG_(printf)("   cline:    splits: 8to4 %,12lu    4to2 %,12lu    2to1 %,12lu\n",
                 stats__cline_64to32splits,
                 stats__cline_32to16splits,
                 stats__cline_16to8splits );
      VG_(printf)("   cline: pulldowns: 8to4 %,12lu    4to2 %,12lu    2to1 %,12lu\n",
                 stats__cline_64to32pulldown,
                 stats__cline_32to16pulldown,
                 stats__cline_16to8pulldown );

      VG_(printf)("\n");
   }
}

static void hg_pre_clo_init ( void )
{
   VG_(details_name)            ("Helgrind");
   VG_(details_version)         (NULL);
   VG_(details_description)     ("a thread error detector");
   VG_(details_copyright_author)(
      "Copyright (C) 2007-2007, and GNU GPL'd, by OpenWorks LLP et al.");
   VG_(details_bug_reports_to)  (VG_BUGS_TO);
   VG_(details_avg_translation_sizeB) ( 200 );

   VG_(basic_tool_funcs)          (hg_post_clo_init,
                                   hg_instrument,
                                   hg_fini);

   VG_(needs_core_errors)         ();
   VG_(needs_tool_errors)         (hg_eq_Error,
                                   hg_pp_Error,
                                   False,/*show TIDs for errors*/
                                   hg_update_extra,
                                   hg_recognised_suppression,
                                   hg_read_extra_suppression_info,
                                   hg_error_matches_suppression,
                                   hg_get_error_name,
                                   hg_print_extra_suppression_info);

   VG_(needs_command_line_options)(hg_process_cmd_line_option,
                                   hg_print_usage,
                                   hg_print_debug_usage);
   VG_(needs_client_requests)     (hg_handle_client_request);

   // FIXME?
   //VG_(needs_sanity_checks)       (hg_cheap_sanity_check,
   //                                hg_expensive_sanity_check);

   VG_(needs_malloc_replacement)  (hg_cli__malloc,
                                   hg_cli____builtin_new,
                                   hg_cli____builtin_vec_new,
                                   hg_cli__memalign,
                                   hg_cli__calloc,
                                   hg_cli__free,
                                   hg_cli____builtin_delete,
                                   hg_cli____builtin_vec_delete,
                                   hg_cli__realloc,
                                   HG_CLI__MALLOC_REDZONE_SZB );

   VG_(needs_data_syms)();

   //VG_(needs_xml_output)          ();

   VG_(track_new_mem_startup)     ( evh__new_mem_w_perms );
   VG_(track_new_mem_stack_signal)( evh__die_mem );
   VG_(track_new_mem_brk)         ( evh__new_mem );
   VG_(track_new_mem_mmap)        ( evh__new_mem_w_perms );
   VG_(track_new_mem_stack)       ( evh__new_mem );

   // FIXME: surely this isn't thread-aware
   VG_(track_copy_mem_remap)      ( shadow_mem_copy_range );

   VG_(track_change_mem_mprotect) ( evh__set_perms );

   VG_(track_die_mem_stack_signal)( evh__die_mem );
   VG_(track_die_mem_brk)         ( evh__die_mem );
   VG_(track_die_mem_munmap)      ( evh__die_mem );
   VG_(track_die_mem_stack)       ( evh__die_mem );

   // FIXME: what is this for?
   VG_(track_ban_mem_stack)       (NULL);

   VG_(track_pre_mem_read)        ( evh__pre_mem_read );
   VG_(track_pre_mem_read_asciiz) ( evh__pre_mem_read_asciiz );
   VG_(track_pre_mem_write)       ( evh__pre_mem_write );
   VG_(track_post_mem_write)      (NULL);

   /////////////////

   VG_(track_pre_thread_ll_create)( evh__pre_thread_ll_create );
   VG_(track_pre_thread_ll_exit)  ( evh__pre_thread_ll_exit );

   VG_(track_start_client_code)( evh__start_client_code );
   VG_(track_stop_client_code)( evh__stop_client_code );

   initialise_data_structures();

   /* Ensure that requirements for "dodgy C-as-C++ style inheritance"
      as described in comments at the top of pub_tool_hashtable.h, are
      met.  Blargh. */
   tl_assert( sizeof(void*) == sizeof(struct _MallocMeta*) );
   tl_assert( sizeof(UWord) == sizeof(Addr) );
   hg_mallocmeta_table
      = VG_(HT_construct)( "hg_malloc_metadata_table" );

   /* a SecMap must contain an integral number of CacheLines */
   tl_assert(0 == (N_SECMAP_ARANGE % N_LINE_ARANGE));
   /* also ... a CacheLine holds an integral number of trees */
   tl_assert(0 == (N_LINE_ARANGE % 8));
}

VG_DETERMINE_INTERFACE_VERSION(hg_pre_clo_init)

/*--------------------------------------------------------------------*/
/*--- end                                                hg_main.c ---*/
/*--------------------------------------------------------------------*/