Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
CMPSC 473 Extended Attributes
Single person project. Do your own work!
In this project, you will extend a provided RAM disk file system with extended attributes that are used to store
ad hoc metadata with files.
The File System
The provided RAM disk file system stores its “disk” configuration in memory when you run the project. That is,
the layout of the disk is exactly the layout in memory. You will extend this file system with extended attribute
that can be added to any file. Fortunately, this extension is largely orthogonal to the provided file system, but
there are some lessons to be learned from studying and understanding the file system code provided that you can
apply.
The file system is defined by its structure shown below.
The file system consists of a series of blocks, which although in memory in your project correspond to the
layout of the file system on disk. The main structures used to implement the different types of file system
metadata, including blocks are defined in the file cmpsc473-filesys.h in the project code. Do not modify this
file.
Each block is prefaced by its block metadata specified in a dblock_t structure. The block metadata determines
the type of block (free, which also indicates whether the type is to be determined), the next block in a list of free
blocks (next), and metadata about the block in a union data structure (st), which may either reference a bitmap
(for a dentry block) or the end of the block data (e.g., for data blocks).
The last field (data[0]) is a reference to the block’s data. Note that although this field is declared as an array of
size 0, it is actually just a reference to a field whose size is determined at runtime. In the case of blocks, the data
in the block is determine by the block type (see BLOCK definitions in cmpsc473-filesys.h). While this may be
uncommon in user-space programs, it is common in the Linux kernel – to avoid wasting memory for objects
whose size is only known at runtime.
Below, we detail the file system structure shown in the figure above.
Block 0 is the file system block or superblock, which defines the overall structure of the file system
using a structure of type dfilesys_t. This block states the number of blocks in the file system (bsize field),
the offset in blocks to the next free block (firstfree field), and the offset to the root directory block of the
file system (root field).
Block 1 stores the root directory block of type ddir_t, which is the only directory in our filesystem. Each
on-disk directory stores a set of references to directory entries (dentries) using a hash table. The on-disk
directory stores the number of buckets in its hashtable (buckets), the free dentry block (see below) for
storing the next directory entry (freeblk), and the first free dentry slot index in that dentry block (free).
The last field (data[0] indicates the start of the hash table (the first bucket) for the directory block.
Block 2 is a directory entry (dentry) block, which stores a series of directory entries (dentries) of type
ddentry_t. The dentries store information about each entry in a directory (i.e., file or subdirectory). In this
project, we only have files, so there are only dentries for files. Each dentry records its file’s file control
block (block), which is the first block for a file, the next dentry in the directory’s hash table (next), and
the file name (name[0]) and name length (name_size).
Thus, directory entry hash table starts in the directory block – by finding the bucket a file name
corresponds to – and then traverses directory entries in dentry blocks using the next pointer to find the
next_dentry block and dentry next_slot using the ddh_t structures. That is probably the most
complicated thing about this file system.
Block 3 is a file control block, which stores the metadata for a particular file using type fcb_t. The file
control block stores the file permissions available to your one process to this file (flags), the size of the
file in bytes (size), the file data blocks (blocks, up to 10 in this project), and the first file attribute block
(attr_block, see Extended Attributes below). There will be several file control blocks – one for each file
created.
Block 4 is a data block, which is used to store file data. The dblock_t structure at the start of every block
is used to manage the file data, which is written starting in the data[0] field, and whose current length is
recorded by the data_end field (in the union).
Extended Attributes
Your task will be to extend the provided file system with extended attributes for files. This section provides
background on extended attributes and your specific project tasks.
Background
Many file systems now support extensible storage of file metadata in the form of extended attributes. Programs
and the kernel may associate attribute-value pairs with a file for any purpose they desire by storing these pairs
with the file as extended attributes.
The way the project uses extended attributes largely follows the extended attribute definitions shown in the
manpages for fsetxattr and for fgetxattr. The full tarball for the project is available here.
This project will focus on two separate data blocks, File Xattr Control Blocks (attribute blocks) and Xattr
Data Blocks (value blocks). These are shown in this diagram.
Block 5 shows a file’s extended attribute (xattr) control block. It is a data block (dblock_t) for one file that
contains a structure describing the xattr information, called the xattr control block of type xcb_t. At the
end of the xcb_t is a reference to an array of attribute structures of type dxattr_t. We will store all the
attribute structures in this one block. The array of 0 indicates that we do not really know how big these
data structures are, nor how many we will really store, so we just leave it to be determined at runtime. This
is actually a common practice in the Linux kernel.
Blocks 6 and 7 are data blocks that store the corresponding attribute values. As values can vary in size,
they are just written like a log to the data block. That is, if we first write an attribute x’s value 10, it is
written to the data block at the beginning. If we change the value to 11, then we write this value after 10,
and update the location in the attribute structure (value_offset). As shown, values can span multiple value
blocks.
Using the extended attributes for a file entails allocating and initializing an xattr control block (on the first
attribute to be set). The xcb_t structure maintains the number of extended attributes stored (no_xattrs)
(one dxattr_t entry in this block for each), the size of the dxattrs for this block (size), and the references to
the value blocks (value_blocks) used for the correpsonding attribute values.
For subsequent attributes, create a new dxattr_t structure (after the current ones) and add the value at the
end of the xattr data block. Make sure that the dxattr_t for the attribute name stores the offset of the value
in the value blocks (in the field value_offset). We can remove a value by setting it to a blank value. We
never remove an attribute once created. Consider leveraging diskWrite to write the value because the
value may span multiple blocks.
In the course of setting attributes, you will have to implement the processing of the flags
XATTR_CREATE and XATTR_REPLACE, where the former requires that no xattr of that name has
been created previous (to prevent collisions) and where the latter requires that the xattr already be defined.
See more below.
Getting an attribute value entails, retrieving the file xattr control block, reading the xcb_t structure to find
the dxattr_t structure with the name string, getting the location of the value string and its size from that
dxattr_t structure, and retrieving the corresponding value from the appropriate block. I create the buffer
for you and print the value. Consider using diskRead as values may span multiple blocks.
Project Tasks
In particular, you are going to be required to implement four functions to enable use of extended attributes.
int fileSetAttr( unsigned int fd, char *name, char *value, unsigned int name_size, unsigned int
value_size, unsigned int flags ): This function sets an attribute of name (length of name_size) of a file
specified by the descriptor fd to value (length of value_size) given the flags value. The flags values can
be XATTR_CREATE, which requires that the attribute not be assigned to the file previously, and
XATTR_REPLACE, which requires that the attribute already has been assigned to the file. Your code
needs to return an error if the conditions are not consistent with the flags. Otherwise, your code should set
the attribute’s value (more detail below).
A key function of your code will be to retrieve a block to store extended attributes for the file and assign it
to the file (file->attr_block). Once assigned, this block should also be made available to the in-memory
file (file_t) and the file control block (fcb_t). Then, the attr_block index can then be retrieved from either
the file or the disk, but you should look for this value on the file structure before reading from the fcb (the
disk). The same should be done for fileGetAttr below.
int fileGetAttr( unsigned int fd, char *name, char *value, unsigned int name_size, unsigned int size
): This function retrieves the value of a file’s attribute name (length of name_size). The function also
takes a buffer for the value, called value, that is allocated to accept string of up to size bytes. Your code
should return the number of bytes read into the value buffer. If no attribute of name is assigned, then
nothing (0 bytes) is returned.
int diskSetAttr( unsigned int attr_block, char *name, char *value, unsigned int name_size, unsigned
int value_size ): Writes the value to a disk data block associating it with the name attribute. As described
in detail below, attr_block is the data block for attribute structures (dxattr_t), so diskSetAttr must create
a structure for name if not already there. The attribute values are stored in separate data blocks referenced
from the attribute structure.
int diskGetAttr( unsigned int attr_block, char *name, char *value, unsigned int name_size,
unsigned int size, unsigned int existsp ): Reads the attribute name from the attr_block to retrieve the
attribute data structure. This structure contains an offset in the value data blocks to enable retrieval of the
value which is written to the value buffer up to length size. If the existsp flag is set, then this function
only returns whether the attribute of name exists (regardless of whether it has a non-null value).
Note that reading and writing attributes bears some resemblance to reading and writing generate disk data
(although the structures for attributes are different (see below). However, you should use file/diskRead
and file/diskWrite for guidance.
In the assignment, an output file (p4-ooutput in the tarball) shows the sequence of commands and responses for
your file system. You will run 5 commands to generate this output: ./cmpsc473-p4 your_fs cmdi >>& p4-
output where cmdi is the ith command for (e.g., cmd1 for the first). This program is deterministic, so your
output should match mine (bug disclaimer here).
NOTE: The functionality for cmd1 and cmd2 are part of the provided file system.