1 @node Project 2--User Programs, Project 3--Virtual Memory, Project 1--Threads, Top
2 @chapter Project 2: User Programs
4 Now that you've worked with Pintos and are familiar with its
5 infrastructure and thread package, it's time to start working on the
6 parts of the system that will allow users to run programs on top of
7 your operating system. The base code already supports loading and
8 running a single user program at a time with little interactivity
9 possible. You will allow multiple programs to be loaded in at once,
10 and to interact with the OS via system calls.
12 You will be working out of the @file{userprog} directory for this
13 assignment. However, you will also be interacting with almost every
14 other part of the code for this assignment. We will describe the
15 relevant parts below. If you are confident in your HW1 code, you can
16 build on top of it. However, if you wish you can start with a fresh
17 copy of the code and re-implement @func{thread_join}, which is the
18 only part of project #1 required for this assignment. Your submission
19 should define @code{THREAD_JOIN_IMPLEMENTED} in @file{constants.h}
20 (@pxref{Conditional Compilation}).
22 Up to now, all of the code you have written for Pintos has been part
23 of the operating system kernel. This means, for example, that all the
24 test code from the last assignment ran as part of the kernel, with
25 full access to privileged parts of the system. Once we start running
26 user programs on top of the operating system, this is no longer true.
27 This project deals with consequences of the change.
29 We allow more than one user program to run at a time. Because user
30 programs are written and compiled to work under the illusion that they
31 have the entire machine, when you load into memory and run more than
32 one process at a time, you must manage things correctly to maintain
35 Before we delve into the details of the new code that you'll be
36 working with, you should probably undo the test cases from project 1.
40 * Using the File System::
41 * How User Programs Work::
42 * Virtual Memory Layout::
43 * Global Requirements::
44 * Problem 2-1 Argument Passing::
45 * Problem 2-2 System Calls::
47 * 80x86 Calling Convention::
54 The easiest way to get an overview of the programming you will be
55 doing is to simply go over each part you'll be working with. In
56 @file{userprog}, you'll find a small number of files, but here is
57 where the bulk of your work will be:
62 Loads ELF binaries and starts processes.
66 A simple manager for 80@var{x} page directories and page tables.
67 Although you probably won't want to modify this code for this project,
68 you may want to call some of its functions. In particular,
69 @func{pagedir_get_page} may be helpful for accessing user memory.
73 Whenever a user process wants to access some kernel functionality, it
74 needs to do so via a system call. This is a skeleton system call
75 handler. Currently, it just prints a message and terminates the user
76 process. In part 2 of this project you will add code to do everything
77 else needed by system calls.
81 When a user process performs a privileged or prohibited operation, it
82 traps into the kernel as an ``exception'' or ``fault.''@footnote{We
83 will treat these terms as synonymous. There is no standard
84 distinction between them, although the Intel processor manuals define
85 them slightly differently on 80@var{x}86.} These files handle
86 exceptions. Currently all exceptions simply print a message and
87 terminate the process. Some, but not all, solutions to project 2
88 require modifying @func{page_fault} in this file.
92 The 80@var{x}86 is a segmented architecture. The Global Descriptor
93 Table (GDT) is a table that describes the segments in use. These
94 files set up the GDT. @strong{You should not need to modify these
95 files for any of the projects.} However, you can read the code if
96 you're interested in how the GDT works.
100 The Task-State Segment (TSS) is used for 80@var{x}86 architectural
101 task switching. Pintos uses the TSS only for switching stacks when a
102 user process enters an interrupt handler, as does Linux. @strong{You
103 should not need to modify these files for any of the projects.}
104 However, you can read the code if you're interested in how the TSS
108 Finally, in @file{lib/kernel}, you might want to use
109 @file{bitmap.[ch]}. A bitmap is basically an array of bits, each of
110 which can be true or false. Bitmaps are typically used to keep track
111 of the usage of a large array of (identical) resources: if resource
112 @var{n} is in use, then bit @var{n} of the bitmap is true. You might
113 find it useful for tracking memory pages, for example.
115 @node Using the File System
116 @section Using the File System
118 You will need to use some file system code for this project. First,
119 user programs are loaded from the file system. Second, many of the
120 system calls you must implement deal with the file system. However,
121 the focus of this project is not on the file system code, so we have
122 provided a simple file system in the @file{filesys} directory. You
123 will want to look over the @file{filesys.h} and @file{file.h}
124 interfaces to understand how to use the file system, and especially
125 its many limitations. @strong{You should not modify the file system
126 code for this project}. Proper use of the file system routines now
127 will make life much easier for project 4, when you improve the file
128 system implementation.
130 You need to be able to create and format simulated disks. The
131 @command{pintos} program provides this functionality with its
132 @option{make-disk} command. From the @file{userprog/build} directory,
133 execute @code{pintos make-disk fs.dsk 2}. This command creates a 2 MB
134 simulated disk named @file{fs.dsk}. (It does not actually start
135 Pintos.) Then format the disk by passing the @option{-f} option to
136 Pintos on the kernel's command line: @code{pintos run -f}.
138 You'll need a way to get files in and out of the simulated file
139 system. The @code{pintos} @option{put} and @option{get} commands are
140 designed for this. To copy @file{@var{file}} into the Pintos file
141 system, use the command @file{pintos put @var{file}}. To copy it to
142 the Pintos file system under the name @file{@var{newname}}, add the
143 new name to the end of the command: @file{pintos put @var{file}
144 @var{newname}}. The commands for copying files out of a VM are
145 similar, but substitute @option{get} for @option{get}.
147 Incidentally, these commands work by passing special options
148 @option{-ci} and @option{-co} on the kernel's command line and copying
149 to and from a special simulated disk named @file{scratch.dsk}. If
150 you're very curious, you can look at the @command{pintos} program as
151 well as @file{filesys/fsutil.c} to learn the implementation details,
152 but it's really not relevant for this project.
154 Here's a summary of how you would create and format a disk, copy the
155 @command{echo} program into the new disk, and then run @command{echo}.
156 It assumes that you've already built the tests in
157 @file{tests/userprog} and that the current directory is
158 @file{userprog/build}:
161 pintos make-disk fs.dsk 2
163 pintos put ../../tests/userprog/echo echo
167 You can delete a file from the Pintos file system using the @option{-r
168 @var{file}} kernel option, e.g.@: @code{pintos run -r @var{file}}.
169 Also, @option{-ls} lists the files in the file system and @option{-p
170 @var{file}} prints a file's contents to the display.
172 @node How User Programs Work
173 @section How User Programs Work
175 Pintos can run normal C programs. In fact, it can run any program you
176 want, provided it's compiled into the proper file format, and uses
177 only the system calls you implement. (For example, @func{malloc}
178 makes use of functionality that isn't provided by any of the syscalls
179 we require you to support.) The only other limitation is that Pintos
180 can't run programs using floating point operations, since it doesn't
181 include the necessary kernel functionality to save and restore the
182 processor's floating-point unit when switching threads. You can look
183 in @file{tests/userprog} directory for some examples.
185 Pintos loads ELF executables, where ELF is an executable format used
186 by Linux, Solaris, and many other Unix and Unix-like systems.
187 Therefore, you can use any compiler and linker that produce
188 80@var{x}86 ELF executables to produce programs for Pintos. We
189 recommend using the tools we provide in the @file{tests/userprog}
190 directory. By default, the @file{Makefile} in this directory will
191 compile the test programs we provide. You can edit the
192 @file{Makefile} to compile your own test programs as well.
194 One thing you should realize immediately is that, until you copy a
195 test program to the emulated disk, Pintos will be unable to do very
196 much useful work. You will also find that you won't be able to do
197 interesting things until you copy a variety of programs to the disk.
198 A useful technique is to create a clean reference disk and copy that
199 over whenever you trash your @file{fs.dsk} beyond a useful state,
200 which may happen occasionally while debugging.
202 @node Virtual Memory Layout
203 @section Virtual Memory Layout
205 Virtual memory in Pintos is divided into two regions: user virtual
206 memory and kernel virtual memory. User virtual memory ranges from
207 virtual address 0 up to @code{PHYS_BASE}, which is defined in
208 @file{threads/mmu.h} and defaults to @t{0xc0000000} (3 GB). Kernel
209 virtual memory occupies the rest of the virtual address space, from
210 @code{PHYS_BASE} up to 4 GB.
212 User virtual memory is per-process. Conceptually, each process is
213 free to use the entire space of user virtual memory however it
214 chooses. When the kernel switches from one process to another, it
215 also switches user virtual address spaces by switching the processor's
216 page directory base register (see @func{pagedir_activate in
217 @file{userprog/pagedir.c}}. @struct{thread} contains a pointer to a
218 process's page directory.
220 Kernel virtual memory is global. It is always mapped the same way,
221 regardless of what user process or kernel thread is running. In
222 Pintos, kernel virtual memory is mapped one-to-one to physical
223 memory. That is, virtual address @code{PHYS_ADDR} accesses physical
224 address 0, virtual address @code{PHYS_ADDR} + @t{0x1234} access
225 physical address @t{0x1234}, and so on up to the size of the machine's
228 User programs can only access user virtual memory. An attempt to
229 access kernel virtual memory will cause a page fault, handled by
230 @func{page_fault} in @file{userprog/exception.c}, and the process
231 will be terminated. Kernel threads can access both kernel virtual
232 memory and, if a user process is running, the user virtual memory of
233 the running process. However, even in the kernel, an attempt to
234 access memory at a user virtual address that doesn't have a page
235 mapped into it will cause a page fault.
237 You must handle memory fragmentation gracefully, that is, a process
238 that needs @var{N} pages of memory must not require that all @var{N}
239 be contiguous. In fact, it must not require that any of the pages be
242 @node Global Requirements
243 @section Global Requirements
245 For testing and grading purposes, we have some simple overall
250 The kernel should print out the program's name and exit status
251 whenever a process terminates, e.g.@: @code{shell: exit(-1)}, whether
252 termination is due to a call to the @code{exit} system call or for
253 another reason. The name printed should be the full name passed to
254 @func{process_execute}, except that it is acceptable to truncate it to
255 15 characters to allow for the limited space in @struct{thread}.
259 Do not print a message when a kernel thread that is not a process
263 Do not print messages about process termination for the @code{halt}
267 No message need be printed when a process that fails to load.
271 Aside from this, the kernel should print out no other messages that
272 Pintos as provided doesn't already print. You
273 may understand all those debug messages, but we won't, and it just
274 clutters our ability to see the stuff we care about.
277 Additionally, while it may be useful to hard-code which process will
278 run at startup while debugging, before you submit your code you must
279 make sure that it takes the start-up process name and arguments from
280 the @samp{-ex} argument. For example, running @code{pintos run -ex
281 "testprogram 1 2 3 4"} will spawn @samp{testprogram 1 2 3 4} as the
285 @node Problem 2-1 Argument Passing
286 @section Problem 2-1: Argument Passing
288 Currently, @func{process_execute} does not support passing arguments
289 to new processes. UNIX and other operating systems do allow passing
290 command line arguments to a program, which accesses them via the argc,
291 argv arguments to main. You must implement this functionality by
292 extending @func{process_execute} so that instead of simply taking a
293 program file name as its argument, it divides it into words at spaces.
294 The first word is the program name, the second word is the first
295 argument, and so on. That is, @code{process_execute("grep foo bar")}
296 should run @command{grep} passing two arguments @code{foo} and
297 @file{bar}. A few details:
301 Multiple spaces are considered the same as a single space, so that
302 @code{process_execute("grep foo bar")} would be equivalent to our
306 You can impose a reasonable limit on the length of the command line
307 arguments. For example, you could limit the arguments to those that
308 will fit in a single page (4 kB).
311 You can parse the argument strings any way you like. If you're lost,
312 look at @func{strtok_r}, prototyped in @file{lib/string.h} and
313 implemented with thorough comments in @file{lib/string.c}. You can
314 find more about it by looking at the man page (run @code{man strtok_r}
318 @xref{80x86 Calling Convention}, for information on exactly how you
319 need to set up the stack.
322 @strong{This functionality is extremely important.} Almost all our
323 test cases rely on being able to pass arguments, so if you don't get
324 this right, a lot of things will not appear to work correctly with our
325 tests. If the tests fail, so do you. Fortunately, this part
326 shouldn't be too hard.
328 @node Problem 2-2 System Calls
329 @section Problem 2-2: System Calls
331 Implement the system call handler in @file{userprog/syscall.c} to
332 properly deal with all the system calls described below. Currently,
333 it ``handles'' system calls by terminating the process. You will need
334 to decipher system call arguments and take the appropriate action for
337 You are required to support the following system calls, whose syscall
338 numbers are defined in @file{lib/syscall-nr.h} and whose C functions
339 called by user programs are prototyped in @file{lib/user/syscall.h}:
343 @itemx void halt (void)
344 Stops Pintos by calling @func{power_off} (declared in
345 @file{threads/init.h}). Note that this should be seldom used, since
346 then you lose some information about possible deadlock situations,
350 @itemx void exit (int @var{status})
351 Terminates the current user program, returning @var{status} to the
352 kernel. If the process's parent @func{join}s it, this is the status
353 that will be returned. Conventionally, a @var{status} of 0 indicates
354 a successful exit. Other values may be used to indicate user-defined
355 conditions (usually errors).
358 @itemx pid_t exec (const char *@var{cmd_line})
359 Runs the executable whose name is given in @var{cmd_line}, passing any
360 given arguments, and returns the new process's program id (pid). If
361 there is an error loading this program, may return pid -1, which
362 otherwise should not be a valid id number.
365 @itemx int join (pid_t @var{pid})
366 Joins the process @var{pid}, using the join rules from the last
367 assignment, and returns the process's exit status. If the process was
368 terminated by the kernel (i.e.@: killed due to an exception), the exit
369 status should be -1. If the process was not a child of the calling
370 process, the return value is undefined (but kernel operation must not
374 @itemx bool create (const char *@var{file}, unsigned @var{initial_size})
375 Create a new file called @var{file} initially @var{initial_size} bytes
376 in size. Returns true if successful, false otherwise.
379 @itemx bool remove (const char *@var{file})
380 Delete the file called @var{file}. Returns true if successful, false
384 @itemx int open (const char *@var{file})
385 Open the file called @var{file}. Returns a nonnegative integer handle
386 called a ``file descriptor'' (fd), or -1 if the file could not be
387 opened. All open files associated with a process should be closed
388 when the process exits or is terminated.
390 File descriptors numbered 0 and 1 are reserved for the console: fd 0
391 is standard input (@code{stdin}), fd 1 is standard output
392 (@code{stdout}). These special file descriptors are valid as system
393 call arguments only as explicitly described below.
396 @itemx int filesize (int @var{fd})
397 Returns the size, in bytes, of the file open as @var{fd}.
400 @itemx int read (int @var{fd}, void *@var{buffer}, unsigned @var{size})
401 Read @var{size} bytes from the file open as @var{fd} into
402 @var{buffer}. Returns the number of bytes actually read (0 at end of
403 file), or -1 if the file could not be read (due to a condition other
404 than end of file). Fd 0 reads from the keyboard using
408 @itemx int write (int @var{fd}, const void *@var{buffer}, unsigned @var{size})
409 Write @var{size} bytes from @var{buffer} to the open file @var{fd}.
410 Returns the number of bytes actually written, or -1 if the file could
411 not be written. Fd 1 writes to the console.
414 @itemx void seek (int @var{fd}, unsigned @var{position})
415 Changes the next byte to be read or written in open file @var{fd} to
416 @var{position}, expressed in bytes from the beginning of the file.
417 (Thus, a @var{position} of 0 is the file's start.)
419 A seek past the current end of a file is not an error. A later read
420 obtains 0 bytes, indicating end of file. A later write extends the
421 file, filling any unwritten gap with zeros. (However, in Pintos files
422 have a fixed length until project 4 is complete, so writes past end of
423 file will return an error.) These semantics are implemented in the
424 file system and do not require any special effort in system call
428 @itemx unsigned tell (int @var{fd})
429 Returns the position of the next byte to be read or written in open
430 file @var{fd}, expressed in bytes from the beginning of the file.
433 @itemx void close (int @var{fd})
434 Close file descriptor @var{fd}.
437 The file defines other syscalls. Ignore them for now. You will
438 implement some of them in project 3 and the rest in project 4, so be
439 sure to design your system with extensibility in mind.
441 To implement syscalls, you will need to provide a way of copying data
442 from the user's virtual address space into the kernel and vice versa.
443 This can be a bit tricky: what if the user provides an invalid
444 pointer, a pointer into kernel memory, or points to a block that is
445 partially in one of those regions? You should handle these cases by
446 terminating the user process. You will need this code before you can
447 even obtain the system call number, because the system call number is
448 on the user's stack in the user's virtual address space. We recommend
449 writing and testing this code before implementing any other system
452 You must make sure that system calls are properly synchronized so that
453 any number of user processes can make them at once. In particular, it
454 is not safe to call into the filesystem code provided in the
455 @file{filesys} directory from multiple threads at once. For now, we
456 recommend adding a single lock that controls access to the filesystem
457 code. You should acquire this lock before calling any functions in
458 the @file{filesys} directory, and release it afterward. Don't forget
459 that @func{process_execute} also accesses files. @strong{For now, we
460 recommend against modifying code in the @file{filesys} directory.}
462 We have provided you a function for each system call in
463 @file{lib/user/syscall.c}. These provide a way for user processes to
464 invoke each system call from a C program. Each of them calls an
465 assembly language routine in @file{lib/user/syscall-stub.S}, which in
466 turn invokes the system call interrupt and returns.
468 When you're done with this part, and forevermore, Pintos should be
469 bulletproof. Nothing that a user program can do should ever cause the
470 OS to crash, halt, assert fail, or otherwise stop running. The sole
471 exception is a call to the @code{halt} system call.
473 If a system call is passed an invalid argument, acceptable options
474 include returning an error value (for those calls that return a
475 value), returning an undefined value, or terminating the process.
477 @xref{System Calls}, for more information on how syscalls work.
479 @node User Programs FAQ
484 @b{Do we need a working project 1 to implement project 2?}
486 You may find the code for @func{thread_join} to be useful in
487 implementing the join syscall, but besides that, you can use
488 the original code provided for project 1.
491 @b{@samp{pintos put} always panics.}
493 Here are the most common causes:
497 The disk hasn't yet been formatted (with @samp{pintos run -f}).
500 The filename specified is too long. The file system limits file names
501 to 14 characters. If you're using a command like @samp{pintos put
502 ../../tests/userprog/echo}, that overflows the limit. Use
503 @samp{pintos put ../../tests/userprog/echo echo} to put the file under
504 the name @file{echo} instead.
507 The file is too big. The file system has a 63 kB limit.
511 @b{All my user programs die with page faults.}
513 This will generally happen if you haven't implemented problem 2-1
514 yet. The reason is that the basic C library for user programs tries
515 to read @var{argc} and @var{argv} off the stack. Because the stack
516 isn't properly set up yet, this causes a page fault.
519 @b{I implemented 2-1 and now all my user programs die with
520 @samp{system call!}.}
522 Every reasonable program tries to make at least one system call
523 (@func{exit}) and most programs make more than that. Notably,
524 @func{printf} invokes the @code{write} system call. The default
525 system call handler just prints @samp{system call!} and terminates the
526 program. You'll have to implement 2-2 before you see anything more
527 interesting. Until then, you can use @func{hex_dump} to convince
528 yourself that 2-1 is implemented correctly (@pxref{Argument Passing to
532 @b{Is there a way I can disassemble user programs?}
534 The @command{i386-elf-objdump} utility can disassemble entire user
535 programs or object files. Invoke it as @code{i386-elf-objdump -d
536 @var{file}}. You can also use @code{i386-elf-gdb}'s
537 @command{disassemble} command to disassemble individual functions in
538 object files compiled with debug information.
541 @b{Why can't I use many C include files in my Pintos programs?}
543 The C library we provide is very limited. It does not include many of
544 the features that are expected of a real operating system's C library.
545 The C library must be built specifically for the operating system (and
546 architecture), since it must make system calls for I/O and memory
547 allocation. (Not all functions do, of course, but usually the library
548 is compiled as a unit.)
551 @b{Can I use lib@var{foo} in my Pintos programs?}
553 The chances are good that lib@var{foo} uses parts of the C library
554 that Pintos doesn't implement. It will probably take at least some
555 porting effort to make it work under Pintos. Notably, the Pintos
556 userland C library does not have a @func{malloc} implementation.
559 @b{How do I compile new user programs?}
561 You need to modify @file{tests/Makefile}.
564 @b{What's the difference between @code{tid_t} and @code{pid_t}?}
566 A @code{tid_t} identifies a kernel thread, which may have a user
567 process running in it (if created with @func{process_execute}) or not
568 (if created with @func{thread_create}). It is a data type used only
571 A @code{pid_t} identifies a user process. It is used by user
572 processes and the kernel in the @code{exec} and @code{join} system
575 You can choose whatever suitable types you like for @code{tid_t} and
576 @code{pid_t}. By default, they're both @code{int}. You can make them
577 a one-to-one mapping, so that the same values in both identify the
578 same process, or you can use a more complex mapping. It's up to you.
581 @b{I can't seem to figure out how to read from and write to user
582 memory. What should I do?}
584 The kernel must treat user memory delicately. As part of a system
585 call, the user can pass to the kernel a null pointer, a pointer to
586 unmapped virtual memory, or a pointer to kernel virtual address space
587 (above @code{PHYS_BASE}). All of these types of invalid pointers must
588 be rejected without harm to the kernel or other running processes. At
589 your option, the kernel may handle invalid pointers by terminating the
590 process or returning from the system call with an error.
592 There are at least two reasonable ways to do this correctly. The
593 first method is to ``verify then access'':@footnote{These terms are
594 made up for this document. They are not standard terminology.} verify
595 the validity of a user-provided pointer, then dereference it. If you
596 choose this route, you'll want to look at the functions in
597 @file{userprog/pagedir.c} and in @file{threads/mmu.h}. This is the
598 simplest way to handle user memory access.
600 The second method is to ``assume and react'': directly dereference
601 user pointers, after checking that they point below @code{PHYS_BASE}.
602 Invalid user pointers will then cause a ``page fault'' that you can
603 handle by modifying the code for @func{page_fault} in
604 @file{userprog/exception.cc}. This technique is normally faster
605 because it takes advantage of the processor's MMU, so it tends to be
606 used in real kernels (including Linux).
608 In either case, you need to make sure not to ``leak'' resources. For
609 example, suppose that your system call has acquired a lock or
610 allocated a page of memory. If you encounter an invalid user pointer
611 afterward, you must still be sure to release the lock or free the page
612 of memory. If you choose to ``verify then access,'' then this should
613 be straightforward, but for ``assume and react'' it's more difficult,
614 because there's no way to return an error code from a memory access.
615 Therefore, for those who want to try the latter technique, we'll
616 provide a little bit of helpful code:
619 /* Tries to copy a byte from user address USRC to kernel address DST.
620 Returns true if successful, false if USRC is invalid. */
621 static inline bool get_user (uint8_t *dst, const uint8_t *usrc) {
623 asm ("movl $1f, %%eax; movb %2, %%al; movb %%al, %0; 1:"
624 : "=m" (*dst), "=&a" (eax) : "m" (*usrc));
628 /* Tries write BYTE to user address UDST.
629 Returns true if successful, false if UDST is invalid. */
630 static inline bool put_user (uint8_t *udst, uint8_t byte) {
632 asm ("movl $1f, %%eax; movb %b2, %0; 1:"
633 : "=m" (*udst), "=&a" (eax) : "r" (byte));
638 Each of these functions assumes that the user address has already been
639 verified to be below @code{PHYS_BASE}. They also assume that you've
640 modified @func{page_fault} so that a page fault in the kernel causes
641 @code{eax} to be set to 0 and its former value copied into @code{eip}.
644 @b{I'm also confused about reading from and writing to the stack. Can
649 Only non-@samp{char} values will have issues when writing them to
650 memory. If a digit is in a string, it is considered a character.
651 However, the value of @code{argc} would be a non-char.
654 You will need to write characters and non-characters into main memory.
657 When you add items to the stack, you will be decrementing the stack
658 pointer. You'll need to decrement the stack pointer before writing to
662 Each character is 1 byte.
666 @b{Why doesn't keyboard input work with @samp{pintos -v}?}
668 Serial input isn't implemented. Don't use @samp{pintos -v} if you
669 want to use the shell or otherwise provide keyboard input.
673 * Problem 2-1 Argument Passing FAQ::
674 * Problem 2-2 System Calls FAQ::
677 @node Problem 2-1 Argument Passing FAQ
678 @subsection Problem 2-1: Argument Passing FAQ
682 @b{Why is the top of the stack at @t{0xc0000000}? Isn't that off the
683 top of user virtual memory? Shouldn't it be @t{0xbfffffff}?}
685 When the processor pushes data on the stack, it decrements the stack
686 pointer first. Thus, the first (4-byte) value pushed on the stack
687 will be at address @t{0xbffffffc}.
689 Also, the stack should always be aligned to a 4-byte boundary, but
690 @t{0xbfffffff} isn't.
693 @b{Is @code{PHYS_BASE} fixed?}
695 No. You should be able to support @code{PHYS_BASE} values that are
696 any multiple of @t{0x10000000} from @t{0x80000000} to @t{0xc0000000},
697 simply via recompilation.
700 @node Problem 2-2 System Calls FAQ
701 @subsection Problem 2-2: System Calls FAQ
705 @b{Can I just cast a pointer to a @struct{file} object to get a
706 unique file descriptor? Can I just cast a @code{struct thread *} to a
707 @code{pid_t}? It's so much simpler that way!}
709 This is a design decision you will have to make for yourself.
710 However, note that most operating systems do distinguish between file
711 descriptors (or pids) and the addresses of their kernel data
712 structures. You might want to give some thought as to why they do so
713 before committing yourself.
716 @b{Can I set a maximum number of open files per process?}
718 From a design standpoint, it would be better not to set an arbitrary
719 maximum. That said, if your design calls for it, you may impose a
720 limit of 128 open files per process (as the Solaris machines here do).
723 @anchor{Removing an Open File}
724 @b{What happens when two (or more) processes have a file open and one of
727 You should copy the standard Unix semantics for files. That is, when
728 a file is removed an process which has a file descriptor for that file
729 may continue to do operations on that descriptor. This means that
730 they can read and write from the file. The file will not have a name,
731 and no other processes will be able to open it, but it will continue
732 to exist until all file descriptors referring to the file are closed
733 or the machine shuts down.
736 @b{I've discovered that some of my user programs need more than one 4
737 kB page of stack space. What should I do?}
739 You may modify the stack setup code to allocate more than one page of
740 stack space for each process.
743 @node 80x86 Calling Convention
744 @section 80@var{x}86 Calling Convention
746 What follows is a quick and dirty discussion of the 80@var{x}86
747 calling convention. Some of the basics should be familiar from CS
748 107, and if you've already taken CS 143 or EE 182, then you should
749 have seen even more of it. I've omitted some of the complexity, since
750 this isn't a class in how function calls work, so don't expect this to
751 be exactly correct in full, gory detail. If you do want all the
752 details, you can refer to @bibref{SysV-i386}.
754 Whenever a function call happens, you need to put the arguments on the
755 call stack for that function, before the code for that function
756 executes, so that the callee has access to those values. The caller
757 has to be responsible for this (be sure you understand why).
758 Therefore, when you compile a program, the assembly code emitted will
759 have in it, before every function call, a bunch of instructions that
760 prepares for the call in whatever manner is conventional for the
761 machine you're working on. This includes saving registers as needed,
762 putting stuff on the stack, saving the location to return to somewhere
763 (so that when the callee finishes, it knows where the caller code is),
764 and some other bookkeeping stuff. Then you do the jump to the
765 callee's code, and it goes along, assuming that the stack and
766 registers are prepared in the appropriate manner. When the callee is
767 done, it looks at the return location as saved earlier, and jumps back
768 to that location. The caller may then have to do some cleanup:
769 clearing arguments and the return value off the stack, restoring
770 registers that were saved before the call, and so on.
772 If you think about it, some of these things should remind you of
775 As an aside, in general, function calls are not cheap. You have to do
776 a bunch of memory writes to prepare the stack, you need to save and
777 restore registers before and after a function call, you need to write
778 the stack pointer, you have a couple of jumps which probably wrecks
779 some of your caches. This is why inlining code can be much faster.
782 * Argument Passing to main::
785 @node Argument Passing to main
786 @subsection Argument Passing to @code{main()}
788 In @func{main}'s case, there is no caller to prepare the stack
789 before it runs. Therefore, the kernel needs to do it. Fortunately,
790 since there's no caller, there are no registers to save, no return
791 address to deal with, etc. The only difficult detail to take care of,
792 after loading the code, is putting the arguments to @func{main} on
795 (The above is a small lie: most compilers will emit code where main
796 isn't strictly speaking the first function. This isn't an important
797 detail. If you want to look into it more, try disassembling a program
798 and looking around a bit. However, you can just act as if
799 @func{main} is the very first function called.)
801 Pintos is written for the 80@var{x}86 architecture. Therefore, we
802 need to adhere to the 80@var{x}86 calling convention. Basically, you
803 put all the arguments on the stack and move the stack pointer
804 appropriately. You also need to insert space for the function's
805 ``return address'': even though the initial function doesn't really
806 have a caller, its stack frame must have the same layout as any other
807 function's. The program will assume that the stack has been laid out
808 this way when it begins running.
810 So, what are the arguments to @func{main}? Just two: an @samp{int}
811 (@code{argc}) and a @samp{char **} (@code{argv}). @code{argv} is an
812 array of strings, and @code{argc} is the number of strings in that
813 array. However, the hard part isn't these two things. The hard part
814 is getting all the individual strings in the right place. As we go
815 through the procedure, let us consider the following example command:
816 @samp{/bin/ls -l foo bar}.
818 The first thing to do is to break the command line into individual
819 strings: @samp{/bin/ls}, @samp{-l}, @samp{foo}, and @samp{bar}. These
820 constitute the arguments of the command, including the program name
821 itself (which belongs in @code{argv[0]}).
823 These individual, null-terminated strings should be placed on the user
824 stack. They may be placed in any order, as you'll see shortly,
825 without affecting how main works, but for simplicity let's assume they
826 are in reverse order (keeping in mind that the stack grows downward on
827 an 80@var{x}86 machine). As we copy the strings onto the stack, we
828 record their (virtual) stack addresses. These addresses will become
829 important when we write the argument vector (two paragraphs down).
831 After we push all of the strings onto the stack, we adjust the stack
832 pointer so that it is word-aligned: that is, we move it down to the
833 next 4-byte boundary. This is required because we will next be
834 placing several words of data on the stack, and they must be aligned
835 to be read correctly. In our example, as you'll see below,
836 the strings start at address @t{0xffed}. One word below that would be
837 at @t{0xffe9}, so we could in theory put the next word on the stack
838 there. However, since the stack pointer should always be
839 word-aligned, we instead leave the stack pointer at @t{0xffe8}.
841 Once we align the stack pointer, we then push the elements of the
842 argument vector, that is, a null pointer, then the addresses of the
843 strings @samp{/bin/ls}, @samp{-l}, @samp{foo}, and @samp{bar}) onto
844 the stack. This must be done in reverse order, such that
845 @code{argv[0]} is at the lowest virtual address, again because the
846 stack is growing downward. (The null pointer pushed first is because
847 @code{argv[argc]} must be a null pointer.) This is because we are now
848 writing the actual array of strings; if we write them in the wrong
849 order, then the strings will be in the wrong order in the array. This
850 is also why, strictly speaking, it doesn't matter what order the
851 strings themselves are placed on the stack: as long as the pointers
852 are in the right order, the strings themselves can really be anywhere.
853 After we finish, we note the stack address of the first element of the
854 argument vector, which is @code{argv} itself.
856 Then we push @code{argv} (that is, the address of the first element of
857 the @code{argv} array) onto the stack, along with the length of the
858 argument vector (@code{argc}, 4 in this example). This must also be
859 done in this order, since @code{argc} is the first argument to
860 @func{main} and therefore is on first (smaller address) on the
861 stack. Finally, we push a fake ``return address'' and leave the stack
862 pointer to point to its location.
864 All this may sound very confusing, so here's a picture which will
865 hopefully clarify what's going on. This represents the state of the
866 stack and the relevant registers right before the beginning of the
867 user program (assuming for this example that the stack bottom is
873 @multitable {@t{0xbfffffff}} {``return address''} {@t{/bin/ls\0}}
874 @item Address @tab Name @tab Data
875 @item @t{0xbffffffc} @tab @code{*argv[3]} @tab @samp{bar\0}
876 @item @t{0xbffffff8} @tab @code{*argv[2]} @tab @samp{foo\0}
877 @item @t{0xbffffff5} @tab @code{*argv[1]} @tab @samp{-l\0}
878 @item @t{0xbfffffed} @tab @code{*argv[0]} @tab @samp{/bin/ls\0}
879 @item @t{0xbfffffec} @tab word-align @tab @samp{\0}
880 @item @t{0xbfffffe8} @tab @code{argv[4]} @tab @t{0}
881 @item @t{0xbfffffe4} @tab @code{argv[3]} @tab @t{0xbffffffc}
882 @item @t{0xbfffffe0} @tab @code{argv[2]} @tab @t{0xbffffff8}
883 @item @t{0xbfffffdc} @tab @code{argv[1]} @tab @t{0xbffffff5}
884 @item @t{0xbfffffd8} @tab @code{argv[0]} @tab @t{0xbfffffed}
885 @item @t{0xbfffffd4} @tab @code{argv} @tab @t{0xbfffffd8}
886 @item @t{0xbfffffd0} @tab @code{argc} @tab 4
887 @item @t{0xbfffffcc} @tab ``return address'' @tab 0
893 In this example, the stack pointer would be initialized to
896 As shown above, your code should start the stack at the very top of
897 the user virtual address space, in the page just below virtual address
898 @code{PHYS_BASE} (defined in @file{threads/mmu.h}).
900 You may find the non-standard @func{hex_dump} function, declared in
901 @file{<stdio.h>}, useful for debugging your argument passing code.
902 Here's what it would show in the above example, given that
903 @code{PHYS_BASE} is @t{0xc0000000}:
906 bfffffc0 00 00 00 00 | ....|
907 bfffffd0 04 00 00 00 d8 ff ff bf-ed ff ff bf f5 ff ff bf |................|
908 bfffffe0 f8 ff ff bf fc ff ff bf-00 00 00 00 00 2f 62 69 |............./bi|
909 bffffff0 6e 2f 6c 73 00 2d 6c 00-66 6f 6f 00 62 61 72 00 |n/ls.-l.foo.bar.|
913 @section System Calls
915 We have already been dealing with one way that the operating system
916 can regain control from a user program: interrupts from timers and I/O
917 devices. These are ``external'' interrupts, because they are caused
918 by entities outside the CPU.
920 The operating system is also called to deal with software exceptions,
921 which are events generated in response to the code. These can be
922 errors such as a page fault or division by zero. However, exceptions
923 are also the means by which a user program can request services
924 (``system calls'') from the operating system.
926 In the 80@var{x}86 architecture, the @samp{int} instruction is the
927 most commonly used means for invoking system calls. This instruction
928 is handled in the same way as other software exceptions. In Pintos,
929 user programs invoke @samp{int $0x30} to make a system call. The
930 system call number and any additional arguments are expected to be
931 pushed on the stack in the normal fashion before invoking the
934 The normal calling convention pushes function arguments on the stack
935 from right to left and the stack grows downward. Thus, when the
936 system call handler @func{syscall_handler} gets control, the system
937 call number is in the 32-bit word at the caller's stack pointer, the
938 first argument is in the 32-bit word at the next higher address, and
939 so on. The caller's stack pointer is accessible to
940 @func{syscall_handler} as the @samp{esp} member of the @code{struct
941 intr_frame} passed to it.
943 Here's an example stack frame for calling a system call numbered 10
944 with three arguments passed as 1, 2, and 3. The stack addresses are
950 @multitable {@t{0xbffffe7c}} {Value}
951 @item Address @tab Value
952 @item @t{0xbffffe7c} @tab 3
953 @item @t{0xbffffe78} @tab 2
954 @item @t{0xbffffe74} @tab 1
955 @item @t{0xbffffe70} @tab 10
961 In this example, the caller's stack pointer would be at
964 The 80@var{x}86 convention for function return values is to place them
965 in the @samp{EAX} register. System calls that return a value can do
966 so by modifying the @samp{eax} member of @struct{intr_frame}.