/ Check-in [9c6b38b9]
Login

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

Overview
Comment:Remove some branches in walTryBeginRead() that were added by check-in [ce5d13c2de] but became unreachable with the addition of logic in check-in [18b26843] that enabled read-only clients to parse the WAL file into a heap-memory WAL-index, thus guaranteeing that the WAL-index header is always available.
Downloads: Tarball | ZIP archive | SQL archive
Timelines: family | ancestors | descendants | both | readonly-wal-recovery
Files: files | file ages | folders
SHA3-256: 9c6b38b9a96c11bdf9db4ea025720a4f49dcb723fa2e2776edc8453bce85c7e3
User & Date: drh 2017-11-13 05:51:37
Context
2017-11-14
03:42
Improvement to a comment. No changes to code. Closed-Leaf check-in: 486949fc user: drh tags: readonly-wal-recovery
2017-11-13
05:51
Remove some branches in walTryBeginRead() that were added by check-in [ce5d13c2de] but became unreachable with the addition of logic in check-in [18b26843] that enabled read-only clients to parse the WAL file into a heap-memory WAL-index, thus guaranteeing that the WAL-index header is always available. check-in: 9c6b38b9 user: drh tags: readonly-wal-recovery
2017-11-11
20:11
In wal.c: improved comments, new assert() and testcase() macros, and replace some magic numbers with appropriate symbolic constants. check-in: 13ec8a77 user: drh tags: readonly-wal-recovery
Changes
Hide Diffs Unified Diffs Show Whitespace Changes Patch

Changes to src/wal.c.

2414
2415
2416
2417
2418
2419
2420



2421
2422
2423
2424
2425
2426
2427
....
2480
2481
2482
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503
2504
2505
2506
2507
2508
2509
2510
  int mxI;                        /* Index of largest aReadMark[] value */
  int i;                          /* Loop counter */
  int rc = SQLITE_OK;             /* Return code  */
  u32 mxFrame;                    /* Wal frame to lock to */

  assert( pWal->readLock<0 );     /* Not currently locked */




  /* Take steps to avoid spinning forever if there is a protocol error.
  **
  ** Circumstances that cause a RETRY should only last for the briefest
  ** instances of time.  No I/O or other system calls are done while the
  ** locks are held, so the locks should not be held for very long. But 
  ** if we are unlucky, another process that is holding a lock might get
  ** paged out or take a page-fault that is time-consuming to resolve, 
................................................................................
    }
    else if( pWal->bShmUnreliable ){
      return walBeginShmUnreliable(pWal, pChanged);
    }
  }

  assert( pWal->nWiData>0 );
  assert( pWal->apWiData[0] || (pWal->readOnly & WAL_SHM_RDONLY) );
  pInfo = pWal->apWiData[0] ? walCkptInfo(pWal) : 0;
  if( !useWal && (pInfo==0 || pInfo->nBackfill==pWal->hdr.mxFrame)
#ifdef SQLITE_ENABLE_SNAPSHOT
   && (pWal->pSnapshot==0 || pWal->hdr.mxFrame==0
     || 0==memcmp(&pWal->hdr, pWal->pSnapshot, sizeof(WalIndexHdr)))
#endif
  ){
    /* The WAL has been completely backfilled (or it is empty).
    ** and can be safely ignored.
    */
    rc = walLockShared(pWal, WAL_READ_LOCK(0));
    walShmBarrier(pWal);
    if( rc==SQLITE_OK ){
      if( pInfo
       && memcmp((void *)walIndexHdr(pWal), &pWal->hdr, sizeof(WalIndexHdr)) 
      ){
        /* It is not safe to allow the reader to continue here if frames
        ** may have been appended to the log before READ_LOCK(0) was obtained.
        ** When holding READ_LOCK(0), the reader ignores the entire log file,
        ** which implies that the database file contains a trustworthy
        ** snapshot. Since holding READ_LOCK(0) prevents a checkpoint from
        ** happening, this is usually correct.
        **







>
>
>







 







|
|
|











<
|
<







2414
2415
2416
2417
2418
2419
2420
2421
2422
2423
2424
2425
2426
2427
2428
2429
2430
....
2483
2484
2485
2486
2487
2488
2489
2490
2491
2492
2493
2494
2495
2496
2497
2498
2499
2500
2501
2502
2503

2504

2505
2506
2507
2508
2509
2510
2511
  int mxI;                        /* Index of largest aReadMark[] value */
  int i;                          /* Loop counter */
  int rc = SQLITE_OK;             /* Return code  */
  u32 mxFrame;                    /* Wal frame to lock to */

  assert( pWal->readLock<0 );     /* Not currently locked */

  /* useWal may only be set for read/write connections */
  assert( (pWal->readOnly & WAL_SHM_RDONLY)==0 || useWal==0 );

  /* Take steps to avoid spinning forever if there is a protocol error.
  **
  ** Circumstances that cause a RETRY should only last for the briefest
  ** instances of time.  No I/O or other system calls are done while the
  ** locks are held, so the locks should not be held for very long. But 
  ** if we are unlucky, another process that is holding a lock might get
  ** paged out or take a page-fault that is time-consuming to resolve, 
................................................................................
    }
    else if( pWal->bShmUnreliable ){
      return walBeginShmUnreliable(pWal, pChanged);
    }
  }

  assert( pWal->nWiData>0 );
  assert( pWal->apWiData[0]!=0 );
  pInfo = walCkptInfo(pWal);
  if( !useWal && pInfo->nBackfill==pWal->hdr.mxFrame
#ifdef SQLITE_ENABLE_SNAPSHOT
   && (pWal->pSnapshot==0 || pWal->hdr.mxFrame==0
     || 0==memcmp(&pWal->hdr, pWal->pSnapshot, sizeof(WalIndexHdr)))
#endif
  ){
    /* The WAL has been completely backfilled (or it is empty).
    ** and can be safely ignored.
    */
    rc = walLockShared(pWal, WAL_READ_LOCK(0));
    walShmBarrier(pWal);
    if( rc==SQLITE_OK ){

      if( memcmp((void *)walIndexHdr(pWal), &pWal->hdr, sizeof(WalIndexHdr)) ){

        /* It is not safe to allow the reader to continue here if frames
        ** may have been appended to the log before READ_LOCK(0) was obtained.
        ** When holding READ_LOCK(0), the reader ignores the entire log file,
        ** which implies that the database file contains a trustworthy
        ** snapshot. Since holding READ_LOCK(0) prevents a checkpoint from
        ** happening, this is usually correct.
        **