summaryrefslogtreecommitdiff
path: root/src/mongo/db/storage/record.h
blob: 15b2e9c236c4ffcaef0c4619ef0fc14aaea5474c (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
// database.h

/**
*    Copyright (C) 2008 10gen Inc.
*
*    This program is free software: you can redistribute it and/or  modify
*    it under the terms of the GNU Affero General Public License, version 3,
*    as published by the Free Software Foundation.
*
*    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 Affero General Public License for more details.
*
*    You should have received a copy of the GNU Affero General Public License
*    along with this program.  If not, see <http://www.gnu.org/licenses/>.
*
*    As a special exception, the copyright holders give permission to link the
*    code of portions of this program with the OpenSSL library under certain
*    conditions as described in each individual source file and distribute
*    linked combinations including the program with the OpenSSL library. You
*    must comply with the GNU Affero General Public License in all respects for
*    all of the code used other than as permitted herein. If you modify file(s)
*    with this exception, you may extend this exception to your version of the
*    file(s), but you are not obligated to do so. If you do not wish to do so,
*    delete this exception statement from your version. If you delete this
*    exception statement from all source files in the program, then also delete
*    it in the license file.
*/

#pragma once

#include "mongo/bson/bsonobjbuilder.h"
#include "mongo/platform/atomic_word.h"
#include "mongo/db/storage/extent.h"

namespace mongo {

    /* Record is a record in a datafile.  DeletedRecord is similar but for deleted space.

    *11:03:20 AM) dm10gen: regarding extentOfs...
    (11:03:42 AM) dm10gen: an extent is a continugous disk area, which contains many Records and DeleteRecords
    (11:03:56 AM) dm10gen: a DiskLoc has two pieces, the fileno and ofs.  (64 bit total)
    (11:04:16 AM) dm10gen: to keep the headesr small, instead of storing a 64 bit ptr to the full extent address, we keep just the offset
    (11:04:29 AM) dm10gen: we can do this as we know the record's address, and it has the same fileNo
    (11:04:33 AM) dm10gen: see class DiskLoc for more info
    (11:04:43 AM) dm10gen: so that is how Record::myExtent() works
    (11:04:53 AM) dm10gen: on an alloc(), when we build a new Record, we must populate its extentOfs then
    */
#pragma pack(1)
    class Record {
    public:
        enum HeaderSizeValue { HeaderSize = 16 };

        int lengthWithHeaders() const {  _accessing(); return _lengthWithHeaders; }
        int& lengthWithHeaders() {  _accessing(); return _lengthWithHeaders; }

        int extentOfs() const { _accessing(); return _extentOfs; }
        int& extentOfs() { _accessing(); return _extentOfs; }

        int nextOfs() const { _accessing(); return _nextOfs; }
        int& nextOfs() { _accessing(); return _nextOfs; }

        int prevOfs() const {  _accessing(); return _prevOfs; }
        int& prevOfs() {  _accessing(); return _prevOfs; }

        const char * data() const { _accessing(); return _data; }
        char * data() { _accessing(); return _data; }

        const char * dataNoThrowing() const { return _data; }
        char * dataNoThrowing() { return _data; }

        int netLength() const { _accessing(); return _netLength(); }

        /* use this when a record is deleted. basically a union with next/prev fields */
        DeletedRecord& asDeleted() { return *((DeletedRecord*) this); }

        // TODO(ERH): remove
        Extent* myExtent(const DiskLoc& myLoc) { return DiskLoc(myLoc.a(), extentOfs() ).ext(); }

        /* get the next record in the namespace, traversing extents as necessary */
        DiskLoc getNext(const DiskLoc& myLoc); // TODO(ERH): remove
        DiskLoc getPrev(const DiskLoc& myLoc); // TODO(ERH): remove

        struct NP {
            int nextOfs;
            int prevOfs;
        };
        NP* np() { return (NP*) &_nextOfs; }

        // ---------------------
        // memory cache
        // ---------------------

        /**
         * touches the data so that is in physical memory
         * @param entireRecrd if false, only the header and first byte is touched
         *                    if true, the entire record is touched
         * */
        void touch( bool entireRecrd = false ) const;

        /**
         * @return if this record is likely in physical memory
         *         its not guaranteed because its possible it gets swapped out in a very unlucky windows
         */
        bool likelyInPhysicalMemory() const ;

        /**
         * tell the cache this Record was accessed
         * @return this, for simple chaining
         */
        Record* accessed();

        static bool likelyInPhysicalMemory( const char* data );

        /**
         * this adds stats about page fault exceptions currently
         * specically how many times we call _accessing where the record is not in memory
         * and how many times we throw a PageFaultException
         */
        static void appendStats( BSONObjBuilder& b );

        static void appendWorkingSetInfo( BSONObjBuilder& b );
    private:

        int _netLength() const { return _lengthWithHeaders - HeaderSize; }

        /**
         * call this when accessing a field which could hit disk
         */
        void _accessing() const;

        int _lengthWithHeaders;
        int _extentOfs;
        int _nextOfs;
        int _prevOfs;

        /** be careful when referencing this that your write intent was correct */
        char _data[4];

    public:
        static bool MemoryTrackingEnabled;

    };
#pragma pack()

    class DeletedRecord {
    public:

        int lengthWithHeaders() const { _accessing(); return _lengthWithHeaders; }
        int& lengthWithHeaders() { _accessing(); return _lengthWithHeaders; }

        int extentOfs() const { _accessing(); return _extentOfs; }
        int& extentOfs() { _accessing(); return _extentOfs; }

        // TODO: we need to not const_cast here but problem is DiskLoc::writing
        DiskLoc& nextDeleted() const { _accessing(); return const_cast<DiskLoc&>(_nextDeleted); }

        DiskLoc myExtentLoc(const DiskLoc& myLoc) const {
            _accessing();
            return DiskLoc(myLoc.a(), _extentOfs);
        }
        Extent* myExtent(const DiskLoc& myLoc) {
            _accessing();
            return DiskLoc(myLoc.a(), _extentOfs).ext();
        }
    private:

        void _accessing() const;

        int _lengthWithHeaders;
        int _extentOfs;
        DiskLoc _nextDeleted;
    };

    struct RecordStats {
        void record( BSONObjBuilder& b );

        AtomicInt64 accessesNotInMemory;
        AtomicInt64 pageFaultExceptionsThrown;
    };

    // ------------------

    inline DiskLoc Record::getNext(const DiskLoc& myLoc) {
        _accessing();
        if ( _nextOfs != DiskLoc::NullOfs ) {
            /* defensive */
            if ( _nextOfs >= 0 && _nextOfs < 10 ) {
                logContext("Assertion failure - Record::getNext() referencing a deleted record?");
                return DiskLoc();
            }

            return DiskLoc(myLoc.a(), _nextOfs);
        }
        Extent *e = myExtent(myLoc);
        while ( 1 ) {
            if ( e->xnext.isNull() )
                return DiskLoc(); // end of table.
            e = e->xnext.ext();
            if ( !e->firstRecord.isNull() )
                break;
            // entire extent could be empty, keep looking
        }
        return e->firstRecord;
    }

    inline DiskLoc Record::getPrev(const DiskLoc& myLoc) {
        _accessing();

        // Check if we still have records on our current extent
        if ( _prevOfs != DiskLoc::NullOfs ) {
            return DiskLoc(myLoc.a(), _prevOfs);
        }

        // Get the current extent
        Extent *e = myExtent(myLoc);
        while ( 1 ) {
            if ( e->xprev.isNull() ) {
                // There are no more extents before this one
                return DiskLoc();
            }

            // Move to the extent before this one
            e = e->xprev.ext();

            if ( !e->lastRecord.isNull() ) {
                // We have found a non empty extent
                break;
            }
        }

        // Return the last record in our new extent
        return e->lastRecord;
    }

}