SQLite

Check-in [c85202baac]
Login

Many hyperlinks are disabled.
Use anonymous login to enable hyperlinks.

Overview
Comment:Rename the PCache1.mxPinned field to n90pct (since it is 90% of nMax) in order to distinguish it from PGroup.mxPinned. Fix the computation of n90pct so that it does not overflow adversely on excessively large cache sizes.
Downloads: Tarball | ZIP archive
Timelines: family | ancestors | descendants | both | trunk
Files: files | file ages | folders
SHA1: c85202baac49b22a4be36b9733d44f01d83edb28
User & Date: drh 2011-01-26 00:07:03.942
Context
2011-01-26
06:13
Changes to exclusive2.test to make it more deterministic. (check-in: 84b0c2bc71 user: dan tags: trunk)
00:07
Rename the PCache1.mxPinned field to n90pct (since it is 90% of nMax) in order to distinguish it from PGroup.mxPinned. Fix the computation of n90pct so that it does not overflow adversely on excessively large cache sizes. (check-in: c85202baac user: drh tags: trunk)
2011-01-25
18:30
Fix new compiler warnings in pcache1.c that were introduced by the recent performance enhancement patches of [e14649301138b684]. (check-in: c17703ec1e user: drh tags: trunk)
Changes
Side-by-Side Diff Ignore Whitespace Patch
Changes to src/pcache1.c.
70
71
72
73
74
75
76
77

78
79
80
81
82
83
84
70
71
72
73
74
75
76

77
78
79
80
81
82
83
84







-
+







  ** The PGroup mutex must be held when accessing nMax.
  */
  PGroup *pGroup;                     /* PGroup this cache belongs to */
  int szPage;                         /* Size of allocated pages in bytes */
  int bPurgeable;                     /* True if cache is purgeable */
  unsigned int nMin;                  /* Minimum number of pages reserved */
  unsigned int nMax;                  /* Configured "cache_size" value */
  unsigned int mxPinned;              /* nMax*9/10 */
  unsigned int n90pct;                /* nMax*9/10 */

  /* Hash table of all pages. The following variables may only be accessed
  ** when the accessor is holding the PGroup mutex.
  */
  unsigned int nRecyclable;           /* Number of pages in the LRU list */
  unsigned int nPage;                 /* Total number of pages in apHash */
  unsigned int nHash;                 /* Number of slots in apHash[] */
595
596
597
598
599
600
601
602

603
604
605
606
607
608
609
595
596
597
598
599
600
601

602
603
604
605
606
607
608
609







-
+







  PCache1 *pCache = (PCache1 *)p;
  if( pCache->bPurgeable ){
    PGroup *pGroup = pCache->pGroup;
    pcache1EnterMutex(pGroup);
    pGroup->nMaxPage += (nMax - pCache->nMax);
    pGroup->mxPinned = pGroup->nMaxPage + 10 - pGroup->nMinPage;
    pCache->nMax = nMax;
    pCache->mxPinned = nMax*9/10;
    pCache->n90pct = pCache->nMax*9/10;
    pcache1EnforceMaxPage(pGroup);
    pcache1LeaveMutex(pGroup);
  }
}

/*
** Implementation of the sqlite3_pcache.xPagecount method. 
705
706
707
708
709
710
711

712
713

714
715
716

717
718
719
720
721
722
723
705
706
707
708
709
710
711
712
713

714
715
716

717
718
719
720
721
722
723
724







+

-
+


-
+







#ifdef SQLITE_MUTEX_OMIT
  pGroup = pCache->pGroup;
#endif


  /* Step 3: Abort if createFlag is 1 but the cache is nearly full */
  nPinned = pCache->nPage - pCache->nRecyclable;
  assert( nPinned>=0 );
  assert( pGroup->mxPinned == pGroup->nMaxPage + 10 - pGroup->nMinPage );
  assert( pCache->mxPinned == pCache->nMax*9/10 );
  assert( pCache->n90pct == pCache->nMax*9/10 );
  if( createFlag==1 && (
        nPinned>=pGroup->mxPinned
     || nPinned>=(int)pCache->mxPinned
     || nPinned>=(int)pCache->n90pct
     || pcache1UnderMemoryPressure(pCache)
  )){
    goto fetch_out;
  }

  if( pCache->nPage>=pCache->nHash && pcache1ResizeHash(pCache) ){
    goto fetch_out;