Fix a bug in bulk load of bitmap files.

Fix a related bug in the bloom code that uses bitmap stores.

4 files changed, 51 insertions, 32 deletions

diff --git a/dist/api_data.py b/dist/api_data.py
index ed4898bda7e..d63d17228a8 100644
--- a/dist/api_data.py
+++ b/dist/api_data.py
@@ -313,7 +313,9 @@ methods = {
 		resident bitmap to be loaded directly into a file by passing a
 		\c WT_ITEM to WT_CURSOR::set_value where the \c size field
 		indicates the number of records in the bitmap (as specified by
-		the file's \c value_format)''',
+		the file's \c value_format). Bulk load bitmap values must end
+		on a byte boundary relative to the bit count - except for the
+		last set of values loaded.''',
 		type='string'),
 	Config('checkpoint', '', r'''
 		the name of a checkpoint to open (the reserved name
diff --git a/src/bloom/bloom.c b/src/bloom/bloom.c
index b53bd3663e0..3ac0329a22f 100644
--- a/src/bloom/bloom.c
+++ b/src/bloom/bloom.c
@@ -180,12 +180,12 @@ __wt_bloom_finalize(WT_BLOOM *bloom)
 
 	/* Add the entries from the array into the table. */
 	for (i = 0; i < bloom->m; i += values.size) {
-		values.data = bloom->bitstring + i;
+		values.data = bloom->bitstring + (i >> 3);
 		/*
 		 * Shave off some bytes for pure paranoia, in case WiredTiger
 		 * reserves some special sizes.
 		 */
-		values.size = (uint32_t)WT_MIN(bloom->m - i, UINT32_MAX - 100);
+		values.size = (uint32_t)WT_MIN(bloom->m - i, UINT32_MAX - 128);
 		c->set_value(c, &values);
 		WT_ERR(c->insert(c));
 	}
diff --git a/src/btree/rec_write.c b/src/btree/rec_write.c
index a23afca13d9..c3d04b7f765 100644
--- a/src/btree/rec_write.c
+++ b/src/btree/rec_write.c
@@ -1452,6 +1452,35 @@ __wt_rec_row_bulk_insert(WT_CURSOR_BULK *cbulk)
 
 #define	WT_FIX_ENTRIES(btree, bytes)	(((bytes) * 8) / (btree)->bitcnt)
 
+static inline int
+__rec_bulk_insert_split_check(WT_CURSOR_BULK  *cbulk)
+{
+	WT_BTREE *btree;
+	WT_RECONCILE *r;
+	WT_SESSION_IMPL *session;
+
+	session = (WT_SESSION_IMPL *)cbulk->cbt.iface.session;
+	r = cbulk->reconcile;
+	btree = session->btree;
+
+	if (cbulk->entry == cbulk->nrecs) {
+		if (cbulk->entry != 0) {
+			/*
+			 * If everything didn't fit, update the counters and
+			 * split.
+			 *
+			 * Boundary: split or write the page.
+			 */
+			__rec_incr(session, r, cbulk->entry,
+			    __bitstr_size(cbulk->entry * btree->bitcnt));
+			WT_RET(__rec_split(session, r));
+		}
+		cbulk->entry = 0;
+		cbulk->nrecs = WT_FIX_ENTRIES(btree, r->space_avail);
+	}
+	return (0);
+}
+
 /*
  * __wt_rec_col_fix_bulk_insert --
  *	Fixed-length column-store bulk insert.
@@ -1464,7 +1493,7 @@ __wt_rec_col_fix_bulk_insert(WT_CURSOR_BULK *cbulk)
 	WT_RECONCILE *r;
 	WT_SESSION_IMPL *session;
 	const uint8_t *data;
-	uint32_t entries, page_entries, page_size;
+	uint32_t entries, offset, page_entries, page_size;
 
 	session = (WT_SESSION_IMPL *)cbulk->cbt.iface.session;
 	r = cbulk->reconcile;
@@ -1472,41 +1501,28 @@ __wt_rec_col_fix_bulk_insert(WT_CURSOR_BULK *cbulk)
 	cursor = &cbulk->cbt.iface;
 
 	if (cbulk->bitmap) {
+		if (((r->recno - 1) * btree->bitcnt) & 0x7) {
+			WT_RET_MSG(session, EINVAL,
+			    "Bulk bitmap load not aligned on a byte boundary");
+			return (EINVAL);
+		}
 		for (data = cursor->value.data, entries = cursor->value.size;
 		    entries > 0;
 		    entries -= page_entries, data += page_size) {
-			page_entries = WT_MIN(entries,
-			    WT_FIX_ENTRIES(btree, r->space_avail));
-			page_size = __bitstr_size(page_entries * btree->bitcnt);
+			WT_RET(__rec_bulk_insert_split_check(cbulk));
 
-			memcpy(r->first_free, data, page_size);
+			page_entries =
+			    WT_MIN(entries, cbulk->nrecs - cbulk->entry);
+			page_size = __bitstr_size(page_entries * btree->bitcnt);
+			offset = __bitstr_size(cbulk->entry * btree->bitcnt);
+			memcpy(r->first_free + offset, data, page_size);
+			cbulk->entry += page_entries;
 			r->recno += page_entries;
-
-			/* Leave the last page for wrapup. */
-			if (entries > page_entries) {
-				__rec_incr(session, r, page_entries, page_size);
-				WT_RET(__rec_split(session, r));
-			} else
-				cbulk->entry = page_entries;
 		}
 		return (0);
 	}
 
-	if (cbulk->entry == cbulk->nrecs) {
-		if (cbulk->entry != 0) {
-			/*
-			 * If everything didn't fit, update the counters and
-			 * split.
-			 *
-			 * Boundary: split or write the page.
-			 */
-			__rec_incr(session, r, cbulk->entry,
-			    __bitstr_size(cbulk->entry * btree->bitcnt));
-			WT_RET(__rec_split(session, r));
-		}
-		cbulk->entry = 0;
-		cbulk->nrecs = WT_FIX_ENTRIES(btree, r->space_avail);
-	}
+	WT_RET(__rec_bulk_insert_split_check(cbulk));
 
 	__bit_setv(r->first_free,
 	    cbulk->entry, btree->bitcnt, ((uint8_t *)cursor->value.data)[0]);
diff --git a/src/include/wiredtiger.in b/src/include/wiredtiger.in
index deccb7ed980..d6a39ddfb89 100644
--- a/src/include/wiredtiger.in
+++ b/src/include/wiredtiger.in
@@ -530,8 +530,9 @@ struct __wt_session {
 	 * stores\, and allows chunks of a memory resident bitmap to be loaded
 	 * directly into a file by passing a \c WT_ITEM to WT_CURSOR::set_value
 	 * where the \c size field indicates the number of records in the bitmap
-	 * (as specified by the file's \c value_format).,a string; default \c
-	 * false.}
+	 * (as specified by the file's \c value_format). Bulk load bitmap values
+	 * must end on a byte boundary relative to the bit count - except for
+	 * the last set of values loaded..,a string; default \c false.}
 	 * @config{checkpoint, the name of a checkpoint to open (the reserved
 	 * name "WiredTigerCheckpoint" opens the most recent internal checkpoint
 	 * taken for the object).  The cursor does not support data