CS372H Spring 2011 Lab 5: Filesystems

Released Wednesday, March 23, 2011
Due Friday, April 8, 2011, 9:00 PM

Introduction

In this lab, you will implement a simple disk-based file system. The file system itself will be implemented in micro-kernel fashion, outside the kernel but within its own user-space environment. Other environments access the file system by making IPC requests to this special file system environment.

Getting Started

Use Git to commit your Lab 4 source, fetch the latest version of the course repository, and then create a local branch called lab5 based on our lab5 branch, origin/lab5:

tig% cd ~/CS372H/lab
tig% git commit -am 'my solution to lab4'
Created commit 734fab7: my solution to lab4
 4 files changed, 42 insertions(+), 9 deletions(-)
tig% git pull
....
tig% git checkout -b lab5 origin/lab5
Branch lab5 set up to track remote branch refs/remotes/origin/lab5.
Switched to a new branch "lab5"
tig% git merge lab4
Merge made by recursive.
 kern/env.c |   42 +++++++++++++++++++
 1 files changed, 42 insertions(+), 0 deletions(-)
tig% 

The main new component for this part of the lab is the file system environment, located in the new fs directory. Scan through all the files in this directory to get a feel for what all is new. Also, there are some new file system-related source files in the user and lib directories, particularly lib/file.c, lib/fd.c, lib/spawn.c, and new global header files inc/fs.h and inc/fd.h. Be sure to scan through all of these files.

You should run the pingpong, primes, and forktree test cases from lab 4 again after merging in the new lab 5 code. You will need to comment out the ENV_CREATE(fs_fs) line in kern/init.c because fs/fs.c tries to do some I/O, which JOS does allow yet. Similarly, temporarily comment out the call to close_all() in lib/exit.c; this function calls subroutines that you will implement later in the lab, and therefore will panic if called. If your lab 4 code doesn't contain any bugs, the test cases should run fine. Don't proceed until they work. Don't forget to un-comment these lines when you start Exercise 1.

If they don't work, use git diff lab4 | more to review all the changes, making sure there isn't any code you wrote for lab4 (or before) missing from lab 5. Make sure that lab 4 still works.

Lab Requirements

As before, do all of the regular exercises described in the lab and at least one challenge problem. (Some challenge problems are more challenging than others, of course!) Additionally, write up brief answers to the questions posed in the lab and a short (e.g., one or two paragraph) description of what you did to solve your chosen challenge problem. If you implement more than one challenge problem, you only need to describe one of them in the write-up, though of course you are welcome to do more. Furthermore, challenge problems are due at the same time as the last part of a lab, so you do not need to have one completed and written up until you are completely done with the lab. Place the write-up in a file called answers.txt (plain text) in the top level of your lab directory before handing in your work.

Please include a header that contains your name, UTCS username, and lab number and make sure the file is named correctly. If you do not, your answer may not be graded. Please include in your answers.txt the write-up for all of the parts of the lab that you have completed so far (so the lab2b write-up should include the lab2a write-up), but do not include the write-up of previous labs. We may print these out and we would rather not waste paper.

Hand-In Procedure

When you are ready to hand in your lab code and write-up, create a file called slack.txt noting how many slack hours you have used for this and each of the previous assignments -- each lab part counts as a different assignment -- as well as how many you have used in total. (This is to help us agree on the number that you have used.) You must include this file even if you didn't use any slack hours. Then run make turnin-parti where i is the part you want to turnin in the lab directory. This will first do a make clean to clean out any .o files and executables, and then create a tar file called lab5i-handin.tar.gz with the entire contents of your lab directory and submit it via the CS turnin utility. If you submit multiple times, we will take the latest submission and count slack hours accordingly.

As before, we will be grading your solutions with a grading program. You can run make grade in the lab directory to test your kernel with the grading program. You may change any of the kernel source and header files you need to in order to complete the lab, but needless to say you must not change or otherwise subvert the grading code.

File system preliminaries

The file system that you will work with is much simpler than most "real" file systems, but it is powerful enough to provide the basic features: creating, reading, writing, and deleting files organized in a hierarchical directory structure.

We are (for the moment anyway) developing only a single-user operating system, which provides protection sufficient to catch bugs but not to protect multiple mutually suspicious users from each other. Our file system therefore does not support the UNIX notions of file ownership or permissions. Our file system also currently does not support hard links, symbolic links, time stamps, or special device files like most UNIX file systems do.

On-Disk File System Structure

Most UNIX file systems divide available disk space into two main types of regions: inode regions and data regions. UNIX file systems assign one inode to each file in the file system; a file's inode holds critical meta-data about the file such as its stat attributes and pointers to its data blocks. The data regions are divided into much larger (typically 8KB or more) data blocks, within which the file system stores file data and directory meta-data. Directory entries contain file names and pointers to inodes; a file is said to be hard-linked if multiple directory entries in the file system refer to that file's inode. Since our file system will not support hard links, we do not need this level of indirection and therefore can make a convenient simplification: our file system will not use inodes at all and instead will simply store all of a file's (or sub-directory's) meta-data within the (one and only) directory entry describing that file.

Both files and directories logically consist of a series of data blocks, which may be scattered throughout the disk much like the pages of an environment's virtual address space can be scattered throughout physical memory. The file system environment hides the details of block layout, presenting interfaces for reading and writing sequences of bytes at arbitrary offsets within files. The file system environment handles all modifications to directories internally as a part of performing actions such as file creation and deletion. Our file system does, however, allow user environments to read directory meta-data directly (e.g., with read and write), which means that user environments can perform directory scanning operations themselves (e.g., to implement the ls program) rather than having to rely on additional special calls to the file system. The disadvantage of this approach to directory scanning, and the reason that most modern UNIX variants discourage it, is that it makes application programs dependent on the format of directory meta-data, making it difficult to change the file system's internal layout without changing or at least recompiling application programs as well.

Sectors and Blocks

Our file system will use a block size of 4096 bytes, conveniently matching the processor's page size.

Superblocks

File systems typically reserve certain disk blocks at "easy-to-find" locations on the disk (such as the very start or the very end) to hold meta-data describing properties of the file system as a whole, such as the block size, disk size, any meta-data required to find the root directory, the time the file system was last mounted, the time the file system was last checked for errors, and so on. These special blocks are called superblocks.

Our file system will have exactly one superblock, which will always be at block 1 on the disk. Its layout is defined by struct Super in inc/fs.h. Block 0 is typically reserved to hold boot loaders and partition tables, so file systems generally do not use the very first disk block. Many "real" file systems maintain multiple superblocks, replicated throughout several widely-spaced regions of the disk, so that if one of them is corrupted or the disk develops a media error in that region, the other superblocks can still be found and used to access the file system.

The Block Bitmap: Managing Free Disk Blocks

In the same way that the kernel must manage the system's physical memory to ensure that a given physical page is used for only one purpose at a time, a file system must manage the blocks of storage on a disk to ensure that a given disk block is used for only one purpose at a time. In pmap.c you keep the Page structures for all free physical pages on a linked list, page_free_list, to keep track of the free physical pages. In file systems it is more common to keep track of free disk blocks using a bitmap rather than a linked list, because a bitmap is more storage-efficient than a linked list and easier to keep consistent. Searching for a free block in a bitmap can take more CPU time than simply removing the first element of a linked list, but for file systems this isn't a problem because the I/O cost of actually accessing the free block after we find it dominates for performance purposes.

To set up a free block bitmap, we reserve a contiguous region of space on the disk large enough to hold one bit for each disk block. For example, since our file system uses 4096-byte blocks, each bitmap block contains 4096*8=32768 bits, or enough bits to describe 32768 disk blocks. In other words, for every 32768 disk blocks the file system uses, we must reserve one disk block for the block bitmap. A given bit in the bitmap is set if the corresponding block is free, and clear if the corresponding block is in use. The block bitmap in our file system always starts at disk block 2, immediately after the superblock. For simplicity we will reserve enough bitmap blocks to hold one bit for each block in the entire disk, including the blocks containing the superblock and the bitmap itself. We will simply make sure that the bitmap bits corresponding to these special, "reserved" areas of the disk are always clear (marked in-use).

File Meta-data

The layout of the meta-data describing a file in our file system is described by struct File in inc/fs.h. This meta-data includes the file's name, size, type (regular file or directory), and pointers to the blocks comprising the file. As mentioned above, we do not have inodes, so this meta-data is stored in a directory entry on disk. Unlike in most "real" file systems, for simplicity we will use this one File structure to represent file meta-data as it appears both on disk and in memory.

The f_direct array in struct File contains space to store the block numbers of the first 10 (NDIRECT) blocks of the file, which we call the file's direct blocks. For small files up to 10*4096 = 40KB in size, this means that the block numbers of all of the file's blocks will fit directly within the File structure itself. For larger files, however, we need a place to hold the rest of the file's block numbers. For any file greater than 40KB in size, therefore, we allocate an additional disk block, called the file's indirect block, to hold up to 4096/4 = 1024 additional block numbers. Our file system therefore allows files to be up to 1034 blocks, or just over four megabytes, in size. To support larger files, "real" file systems typically support double- and triple-indirect blocks as well.

Directories versus Regular Files

A File structure in our file system can represent either a regular file or a directory; these two types of "files" are distinguished by the type field in the File structure. The file system manages regular files and directory-files in exactly the same way, except that it does not interpret the contents of the data blocks associated with regular files at all, whereas the file system interprets the contents of a directory-file as a series of File structures describing the files and subdirectories within the directory.

The superblock in our file system contains a File structure (the root field in struct Super) that holds the meta-data for the file system's root directory. The contents of this directory-file is a sequence of File structures describing the files and directories located within the root directory of the file system. Any subdirectories in the root directory may in turn contain more File structures representing sub-subdirectories, and so on.

The File System

The goal for this lab is not to have you implement the entire file system, but for you to implement only certain key components. In particular, you will be responsible for reading blocks into the block cache and flushing them back to disk; allocating disk blocks; mapping file offsets to disk blocks; and implementing read, write, and open in the IPC interface. Because you will not be implementing all of the file system yourself, it is very important that you familiarize yourself with the provided code and the various file system interfaces.

Much of the code for the file system has been rewritten for this year, so please let us know if you come across any bugs.

Disk Access

The file system environment in our operating system needs to be able to access the disk, but we have not yet implemented any disk access functionality in our kernel. Instead of taking the conventional "monolithic" operating system strategy of adding an IDE disk driver to the kernel along with the necessary system calls to allow the file system to access it, we will instead implement the IDE disk driver as part of the user-level file system environment. We will still need to modify the kernel slightly, in order to set things up so that the file system environment has the privileges it needs to implement disk access itself.

It is easy to implement disk access in user space this way as long as we rely on polling, "programmed I/O" (PIO)-based disk access and do not use disk interrupts. It is possible to implement interrupt-driven device drivers in user mode as well (the L3 and L4 kernels do this, for example), but it is more difficult since the kernel must field device interrupts and dispatch them to the correct user-mode environment.

The x86 processor uses the IOPL bits in the EFLAGS register to determine whether protected-mode code is allowed to perform special device I/O instructions such as the IN and OUT instructions. Since all of the IDE disk registers we need to access are located in the x86's I/O space rather than being memory-mapped, giving "I/O privilege" to the file system environment is the only thing we need to do in order to allow the file system to access these registers. In effect, the IOPL bits in the EFLAGS register provides the kernel with a simple "all-or-nothing" method of controlling whether user-mode code can access I/O space. In our case, we want the file system environment to be able to access I/O space, but we do not want any other environments to be able to access I/O space at all.

To keep things simple, from now on we will arrange things so that the file system is always user environment 1. (Recall that the idle loop is always user environment 0.)

In the tests that follow, if you fail a test, obj/fs/fs.img is likely to be left inconsistent. Be sure to remove it before running make grade or make qemu.

Do you have to do anything else to ensure that this I/O privilege setting is saved and restored properly when you subsequently switch from one environment to another? Make sure you understand how this environment state is handled.

Read through the files in the new fs directory in the source tree. The file fs/ide.c implements our minimal PIO-based disk driver. The file fs/serv.c contains the umain function for the file system environment.

Note that the GNUmakefile file in this lab sets up QEMU to use the file obj/kern/kernel.img as the image for disk 0 (typically "Drive C" under DOS/Windows) as before, and to use the (new) file obj/fs/fs.img as the image for disk 1 ("Drive D"). In this lab your file system should only ever touch disk 1; disk 0 is used only to boot the kernel. If you manage to corrupt either disk image in some way, you can reset both of them to their original, "pristine" versions simply by typing:

$ rm obj/kern/kernel.img obj/fs/fs.img
$ make

or by doing:

$ make clean
$ make

The Block Cache

In our file system, we will implement a simple "buffer cache" (really just a block cache) with the help of the processor's virtual memory system. The code for the block cache is in fs/bc.c.

Our file system will be limited to handling disks of size 3GB or less. We reserve a large, fixed 3GB region of the file system environment's address space, from 0x10000000 (DISKMAP) up to 0xD0000000 (DISKMAP+DISKMAX), as a "memory mapped" version of the disk. For example, disk block 0 is mapped at virtual address 0x10000000, disk block 1 is mapped at virtual address 0x10001000, and so on. The diskaddr function in fs/bc.c implements this translation from disk block numbers to virtual addresses (along with some sanity checking).

Since our file system environment has its own virtual address space independent of the virtual address spaces of all other environments in the system, and the only thing the file system environment needs to do is to implement file access, it is reasonable to reserve most of the file system environment's address space in this way. It would be awkward for a real file system implementation on a 32-bit machine to do this since modern disks are larger than 3GB. Such a buffer cache management approach may still be reasonable on a machine with a 64-bit address space, such as Intel's Itanium or AMD's Athlon 64 processors.

Of course, it would be unreasonable to read the entire disk into memory, so instead we'll implement a form of demand paging, wherein we only allocate pages in the disk map region and read the corresponding block from the disk in response to a page fault in this region. This way, we can pretend that the entire disk is in memory.

The Block Bitmap

After fs_init sets the bitmap pointer, we can treat bitmap as a packed array of bits, one for each block on the disk. See, for example, block_is_free, which simply checks whether a given block is marked free in the bitmap.

File Operations

We have provided a variety of functions in fs/fs.c to implement the basic facilities you will need to interpret and manage File structures, scan and manage the entries of directory-files, and walk the file system from the root to resolve an absolute pathname. Read through all of the code in fs/fs.c carefully and make sure you understand what each function does before proceeding.

file_block_walk and file_get_block are the workhorses of the file system. For example, file_read and file_write are little more than the bookkeeping atop file_get_block necessary to copy bytes between scattered blocks and a sequential buffer.

Client/Server File System Access

Now that we have implemented the necessary functionality within the file system environment itself, we must make it accessible to other environments that wish to use the file system. Since other environments can't directly call functions in the file system environment, we'll expose access to the file system environment via a remote procedure call, or RPC, abstraction, built atop JOS's IPC mechanism.

Graphically, here's what a call to the file system server (say, read) looks like

      Regular env           FS env
   +---------------+   +---------------+
   |      read     |   |   file_read   |
   |   (lib/fd.c)  |   |   (fs/fs.c)   |
...|.......|.......|...|.......^.......|...............
   |       v       |   |       |       | RPC mechanism
   |  devfile_read |   |  serve_read   |
   |  (lib/file.c) |   |  (fs/serv.c)  |
   |       |       |   |       ^       |
   |       v       |   |       |       |
   |     fsipc     |   |     serve     |
   |  (lib/file.c) |   |  (fs/serv.c)  |
   |       |       |   |       ^       |
   |       v       |   |       |       |
   |   ipc_send    |   |   ipc_recv    |
   |       |       |   |       ^       |
   +-------|-------+   +-------|-------+
           |                   |
           +-------------------+

Everything below the dotted line is simply the mechanics of getting a read request from the regular environment to the file system environment. Starting at the beginning, read (which we provide) works on any file descriptor and simply dispatches to the appropriate device read function, in this case devfile_read (we'll have more device types in future labs, like pipes and network sockets). devfile_read implements read specifically for on-disk files. This and the other devfile_* functions in lib/file.c implement the client side of the FS operations and all work in roughly the same way, bundling up arguments in a request structure, calling fsipc to send the IPC request, and unpacking and returning the results. The fsipc function simply handles the common details of sending a request to the server and receiving the reply.

The file system server code can be found in fs/serv.c. It loops in the serve function, endlessly receiving a request over IPC, dispatching that request to the appropriate handler function, and sending the result back via IPC. In the read example, serve will dispatch to serve_read, which will take care of the IPC details specific to read requests such as unpacking the request structure and finally call file_read to actually perform the file read. The corresponding client code can be found in lib/file.c. Here, the fsipc function handles the common details of sending some request to the server and receiving the reply. The remaining functions wrap the available RPC's as regular C functions that handle bundling up arguments in the request structure, calling fsipc, and unpacking and returning the results.

Recall that JOS's IPC mechanism lets an environment send a single 32-bit number and, optionally, share a page. To send a request from the client to the server, we use the 32-bit number for the request type (the file system server RPCs are numbered, just like how syscalls were numbered) and store the arguments to the request in a union Fsipc on the page shared via the IPC. On the client side, we always share the page at fsipcbuf; on the server side, we map the incoming request page at fsreq (0x0ffff000).

The server also sends the response back via IPC. We use the 32-bit number for the function's return code. For most RPCs, this is all they return. FSREQ_READ and FSREQ_STAT also return data, which they simply write to the page that the client sent its request on. There's no need to send this page in the response IPC, since the client shared it with the file system server in the first place. Also, in its response, FSREQ_OPEN shares with the client a new "Fd page". We'll return to this RPC shortly.

Client-Side File Operations

The functions in lib/file.c are specific to on-disk files, but UNIX file descriptors are a more general notion that also encompasses pipes, console I/O, etc. In JOS, each of these device types has (or will have in later labs) a corresponding struct Dev, with pointers to the functions that implement read/write/etc. for that device type. lib/fd.c implements the general UNIX-like file descriptor interface on top of this. Each struct Fd indicates its device type, and most of the functions in lib/fd.c simply dispatch operations to functions in the appropriate struct Dev.

lib/fd.c also maintains the file descriptor table region in each application environment's address space, starting at FDTABLE. This area reserves a page's worth (4KB) of address space for each of the up to MAXFD (currently 32) file descriptors that the application can have open at once. At any given time, a particular file descriptor table page is mapped if and only if the corresponding file descriptor is in use. Each file descriptor also has an optional "data page" in the region starting at FILEDATA, though we won't use that until later labs.

For nearly all interactions with files, user code should go through the functions in lib/fd.c. The one "public" function in lib/file.c is open, which constructs a new file descriptor by opening a named on-disk file.

Spawning Processes

We have given you the code for spawn which creates a new environment, loads a program image from the file system into it, and then starts the child environment running this program. The parent process then continues running independently of the child. The spawn function effectively acts like a fork in UNIX followed by an immediate exec in the child process.

We implemented spawn rather than a UNIX-style exec because spawn is easier to implement from user space in "exokernel fashion", without special help from the kernel. Think about what you would have to do in order to implement exec in user space, and be sure you understand why it is harder.