120
|
1 =====================================
|
|
2 The MSF File Format
|
|
3 =====================================
|
|
4
|
|
5 .. contents::
|
|
6 :local:
|
|
7
|
|
8 .. _msf_superblock:
|
|
9
|
|
10 The Superblock
|
|
11 ==============
|
|
12 At file offset 0 in an MSF file is the MSF *SuperBlock*, which is laid out as
|
|
13 follows:
|
|
14
|
|
15 .. code-block:: c++
|
|
16
|
|
17 struct SuperBlock {
|
|
18 char FileMagic[sizeof(Magic)];
|
|
19 ulittle32_t BlockSize;
|
|
20 ulittle32_t FreeBlockMapBlock;
|
|
21 ulittle32_t NumBlocks;
|
|
22 ulittle32_t NumDirectoryBytes;
|
|
23 ulittle32_t Unknown;
|
|
24 ulittle32_t BlockMapAddr;
|
|
25 };
|
|
26
|
|
27 - **FileMagic** - Must be equal to ``"Microsoft C / C++ MSF 7.00\\r\\n"``
|
|
28 followed by the bytes ``1A 44 53 00 00 00``.
|
|
29 - **BlockSize** - The block size of the internal file system. Valid values are
|
|
30 512, 1024, 2048, and 4096 bytes. Certain aspects of the MSF file layout vary
|
|
31 depending on the block sizes. For the purposes of LLVM, we handle only block
|
|
32 sizes of 4KiB, and all further discussion assumes a block size of 4KiB.
|
|
33 - **FreeBlockMapBlock** - The index of a block within the file, at which begins
|
|
34 a bitfield representing the set of all blocks within the file which are "free"
|
|
35 (i.e. the data within that block is not used). This bitfield is spread across
|
|
36 the MSF file at ``BlockSize`` intervals.
|
|
37 **Important**: ``FreeBlockMapBlock`` can only be ``1`` or ``2``! This field
|
|
38 is designed to support incremental and atomic updates of the underlying MSF
|
|
39 file. While writing to an MSF file, if the value of this field is `1`, you
|
|
40 can write your new modified bitfield to page 2, and vice versa. Only when
|
|
41 you commit the file to disk do you need to swap the value in the SuperBlock
|
|
42 to point to the new ``FreeBlockMapBlock``.
|
|
43 - **NumBlocks** - The total number of blocks in the file. ``NumBlocks * BlockSize``
|
|
44 should equal the size of the file on disk.
|
|
45 - **NumDirectoryBytes** - The size of the stream directory, in bytes. The stream
|
|
46 directory contains information about each stream's size and the set of blocks
|
|
47 that it occupies. It will be described in more detail later.
|
|
48 - **BlockMapAddr** - The index of a block within the MSF file. At this block is
|
|
49 an array of ``ulittle32_t``'s listing the blocks that the stream directory
|
|
50 resides on. For large MSF files, the stream directory (which describes the
|
|
51 block layout of each stream) may not fit entirely on a single block. As a
|
|
52 result, this extra layer of indirection is introduced, whereby this block
|
|
53 contains the list of blocks that the stream directory occupies, and the stream
|
|
54 directory itself can be stitched together accordingly. The number of
|
|
55 ``ulittle32_t``'s in this array is given by ``ceil(NumDirectoryBytes / BlockSize)``.
|
|
56
|
|
57 The Stream Directory
|
|
58 ====================
|
|
59 The Stream Directory is the root of all access to the other streams in an MSF
|
|
60 file. Beginning at byte 0 of the stream directory is the following structure:
|
|
61
|
|
62 .. code-block:: c++
|
|
63
|
|
64 struct StreamDirectory {
|
|
65 ulittle32_t NumStreams;
|
|
66 ulittle32_t StreamSizes[NumStreams];
|
|
67 ulittle32_t StreamBlocks[NumStreams][];
|
|
68 };
|
|
69
|
|
70 And this structure occupies exactly ``SuperBlock->NumDirectoryBytes`` bytes.
|
|
71 Note that each of the last two arrays is of variable length, and in particular
|
|
72 that the second array is jagged.
|
|
73
|
|
74 **Example:** Suppose a hypothetical PDB file with a 4KiB block size, and 4
|
|
75 streams of lengths {1000 bytes, 8000 bytes, 16000 bytes, 9000 bytes}.
|
|
76
|
|
77 Stream 0: ceil(1000 / 4096) = 1 block
|
|
78
|
|
79 Stream 1: ceil(8000 / 4096) = 2 blocks
|
|
80
|
|
81 Stream 2: ceil(16000 / 4096) = 4 blocks
|
|
82
|
|
83 Stream 3: ceil(9000 / 4096) = 3 blocks
|
|
84
|
|
85 In total, 10 blocks are used. Let's see what the stream directory might look
|
|
86 like:
|
|
87
|
|
88 .. code-block:: c++
|
|
89
|
|
90 struct StreamDirectory {
|
|
91 ulittle32_t NumStreams = 4;
|
|
92 ulittle32_t StreamSizes[] = {1000, 8000, 16000, 9000};
|
|
93 ulittle32_t StreamBlocks[][] = {
|
|
94 {4},
|
|
95 {5, 6},
|
|
96 {11, 9, 7, 8},
|
|
97 {10, 15, 12}
|
|
98 };
|
|
99 };
|
|
100
|
|
101 In total, this occupies ``15 * 4 = 60`` bytes, so ``SuperBlock->NumDirectoryBytes``
|
|
102 would equal ``60``, and ``SuperBlock->BlockMapAddr`` would be an array of one
|
|
103 ``ulittle32_t``, since ``60 <= SuperBlock->BlockSize``.
|
|
104
|
|
105 Note also that the streams are discontiguous, and that part of stream 3 is in the
|
|
106 middle of part of stream 2. You cannot assume anything about the layout of the
|
|
107 blocks!
|
|
108
|
|
109 Alignment and Block Boundaries
|
|
110 ==============================
|
|
111 As may be clear by now, it is possible for a single field (whether it be a high
|
|
112 level record, a long string field, or even a single ``uint16``) to begin and
|
|
113 end in separate blocks. For example, if the block size is 4096 bytes, and a
|
|
114 ``uint16`` field begins at the last byte of the current block, then it would
|
|
115 need to end on the first byte of the next block. Since blocks are not
|
|
116 necessarily contiguously laid out in the file, this means that both the consumer
|
|
117 and the producer of an MSF file must be prepared to split data apart
|
|
118 accordingly. In the aforementioned example, the high byte of the ``uint16``
|
|
119 would be written to the last byte of block N, and the low byte would be written
|
|
120 to the first byte of block N+1, which could be tens of thousands of bytes later
|
|
121 (or even earlier!) in the file, depending on what the stream directory says.
|