Lines Matching full:of

2 .\"	The Regents of the University of California.  All rights reserved.
7 .\" 1. Redistributions of source code must retain the above copyright
8 .\" notice, this list of conditions and the following disclaimer.
10 .\" notice, this list of conditions and the following disclaimer in the
12 .\" 3. Neither the name of the University nor the names of its contributors
18 .\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
21 .\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22 .\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23 .\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25 .\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
36 located at the beginning of the file system's disk partition.
48 bytes with only two levels of indirection,
49 the minimum size of a file system block is 4096 bytes.
50 The size of file system blocks can be any power of two
52 The block size of a file system is recorded in the
63 A cylinder group is comprised of one or more consecutive
66 that includes a redundant copy of the super-block,
69 and summary information describing the usage of data blocks
71 The bit map of available blocks in the cylinder group replaces
73 For each cylinder group a static number of inodes
76 bytes of space in the cylinder group, expecting this
80 placed at the beginning of each cylinder group.
84 could cause the loss of all redundant copies of the super-block.
86 begins at a varying offset from the beginning of the cylinder group.
88 about one track further from the beginning of the cylinder group
92 or platter can be lost without losing all copies of the super-block.
94 the space between the beginning of the cylinder group
95 and the beginning of the cylinder group information
101 with blocks sizes of 16 kilobytes or greater.
102 This is because of a requirement that the first 8 kilobytes of the disk
107 would have to leave an empty space between the end of
108 the boot block and the beginning of the cylinder group.
109 Without knowing the size of the file system blocks,
111 to find the beginning of the first cylinder group.
119 composed of 4096 byte data blocks.
120 In the old file system this file would be composed of 1024 byte blocks.
130 file systems are composed of many small files.
132 Table 1 shows the effect of file system
133 block size on the amount of wasted space in the file system.
135 one of our time sharing
136 systems that has roughly 1.2 gigabytes of on-line storage.
138 about 920 megabytes of formatted space.
154 Table 1 \- Amount of wasted space as a function of block size.
157 The space wasted is calculated to be the percentage of space
171 2, 4, or 8 fragments, each of which is addressable.
172 The lower bound on the size of these fragments is constrained
179 Figure 1 shows a piece of a map from a 4096/1024 file system.
189 Figure 1 \- Example layout of blocks and fragments in a 4096/1024 file system.
192 Each bit in the map records the status of a fragment;
198 Fragments of adjoining blocks cannot be used as a full block,
204 On a file system with a block size of 4096 bytes
205 and a fragment size of 1024 bytes,
206 a file is represented by zero or more 4096 byte blocks of data,
209 space for a small amount of data,
210 the remaining fragments of the block are made
215 three fragment portion of another block.
225 the size of the file has increased*.
227 * A program may be overwriting data in the middle of an existing file
231 one of three conditions exists:
242 If the remainder of the new data contains more than
243 a full block of data, a full block is allocated and
244 the first full block of new data is written there.
255 If the size of the new data plus the size of the data
256 already in the fragments exceeds the size of a full block,
258 The contents of the fragments are copied
259 to the beginning of the block
260 and the remainder of the block is filled with new data.
266 The contents of the existing fragments
275 except for a partial block at the end of the file.
280 For files the optimal size is the block size of the file system
292 The amount of wasted space in the 4096/1024 byte new file system
296 has about the same amount of wasted space as the 512 byte
301 large files and the same amount of space
304 more space for keeping track of available free blocks.
313 gives the minimum acceptable percentage of file system
315 If the number of free blocks drops below this level
317 The value of this parameter may be changed at any time,
320 systems kept less than 90% full (a reserve of 10%).
321 If the number of free blocks falls to zero,
323 because of the inability of the file system to localize
327 files until the amount of free space once again
329 Access rates for files created during periods of little
333 percentage of waste when comparing the organizations given
335 Thus, the percentage of waste in
345 Except for the initial creation of the free list,
346 the old file system ignores the parameters of the underlying hardware.
350 A goal of the new file system is to parameterize the
355 Parameters used include the speed of the processor,
357 and the characteristics of the mass storage devices.
362 adapted to the characteristics of the disk on which
377 without suffering lost time because of an intervening disk revolution.
383 speed of the main processor.
385 The physical characteristics of each disk include
386 the number of blocks per track and the rate at which
389 the number of milliseconds required to skip over a block.
390 The characteristics of the processor include
394 the allocation routines calculate the number of blocks to
397 amount of time that it takes to start a new
399 For programs that sequentially access large amounts of data,
400 this strategy minimizes the amount of time spent waiting for
403 To ease the calculation of finding rotationally optimal blocks,
405 a count of the available blocks in a cylinder
408 so the resolution of the
411 The super-block contains a vector of lists called
414 Each component of the vector
420 It then uses the index of the rotational position to find the appropriate
422 only the relevant parts of the block map to find a free block.
425 minimum number of milliseconds between the completion of a data
426 transfer and the initiation of
431 a rotational separation of 2 milliseconds,
434 the throughput will drop precipitously because of lost disk revolutions
442 characteristics of the new host.
449 the placement of new inodes and data blocks.
451 placement of new directories and files.
461 the locality of reference to minimize seek latency
463 to improve the layout of data to make larger transfers possible
471 the local cylinder group may run out of space
474 total localization can result in a single huge cluster of data
477 goals of localizing data that is concurrently accessed
482 Inodes of files in the same directory are frequently accessed together.
485 The layout policy tries to place all the inodes of
490 than average number of free inodes,
491 and the smallest number of directories already in it.
492 The intent of this policy is to allow the inode clustering policy
493 to succeed most of the time.
494 The allocation of inodes within a cylinder group is done using a
501 This puts a small and constant upper bound on the number of
519 Ideally none of the cylinder groups should ever become completely full.
530 The other spillover points are chosen with the intent of
532 file has used about 25% of the data blocks in a cylinder group.
534 to work well in minimizing the number of completely filled
538 groups that have a greater than average number of free blocks left.
540 a megabyte of data is typically accessible before
542 and the cost of one long seek per megabyte is small.
549 allocates a free block of the requested size that is
555 thus the implementation of the global layout policy
579 Quadratic hash is used because of its speed in finding
586 the most important characteristic of