2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 1/13

ASST2: System Calls 1. Introduction In this assignment you will add process and system call support to your OS/161  kernel. Currently no support exists for running user processes—the tests you have run up to this point have run in the kernel as kernel threads. By the time you finish ASST2 you will have the ability to launch a simple shell and enter a somewhat­familiar UNIX environment. Indeed, future tests will be run as user

processes, not from the kernel menu. In contrast to ASST1, ASST2 requires a great deal more thought and planning. You will be writing a lot more code—while our ASST1 solution adds 624 non­commenting lines of C code 1 , the ASST2 solution adds an additional 2274. We are not giving you many examples. We have not designed your data structures for you. You will need to determine what needs to be implemented, where to put the code required, how the data structures are designed and implemented, and how everything fits together.

We just care that your system calls meet the interface specification. As a final piece of advice, ASST2 and ASST3 begin to produce code bases that are large, complex, and potentially very difficult to debug. You do not want to introduce bugs into your kernel because they

will be very hard to remove. Our advice—slow down, design, think, design again, discuss with your partner, and slow down again. Then write some code. 1.1. Objectives After completing ASST2 you should:

1. Be able to write code that meets a specified interface definition. 2. Understand how to represent processes in an operating system. 3. Have designed and implemented data structures to manage processes in an operating system.

4. Understand how to implement system calls. 1.2. Collaboration Guidelines ASST2 is the first large ops‐class.org assignment. Here are the guidelines for how to work with other

students and with your partner (if you have one): Pair programming to complete the implementation tasks is strongly encouraged. Writing a design document with your partner is strongly encouraged.   2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 2/13

Having one partner work on the file system system calls while the other partner works on process support is a good division of labor. The partner working on the file system system calls may finish first, at which point they can help and continue testing.

Answering the code reading questions side­by­side with your partner is strongly encouraged. Discussing the code reading questions and browsing the source tree with other students is encouraged.

Dividing the code reading questions and development tasks between partners is discouraged.

Any arrangement that results in one partner writing the entire design document is cheating. Any arrangement that results in one partner writing all or almost all of the code is cheating. Copying any answers from anyone who is not your partner or anywhere else and submitting them as your own is cheating. You may not refer to or incorporate any external sources without explicit permission 2 .

2. Assignment Organization Your current OS/161 system has minimal support for running user binaries—nothing that could be considered a true process. ASST2 starts the transformation of OS/161 into a

true multi­tasking operating system.         2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 3/13

After the next assignment, it will be capable of running multiple processes at once from actual compiled programs stored in your account. These programs will be loaded into OS/161 and executed in user mode by System/161. This will occur under the control of your kernel and the command shell in bin/sh. The first step is to read and understand the parts of the system that we have written for you. Our code

can run one user­level C program at a time as long as it doesn't want to do anything but shut the system down. We have provided sample user programs that do this (sbin/{reboot,halt,poweroff), as well as others that make use of features you will be adding in this and future assignments. 2.1. System Call Interface First, however, you must implement the interface between user­mode programs—or user land—and the

kernel. As usual, we provide part of the code you will need. Your job is to identify, design and build the missing pieces. So far, all the code you have written for OS/161 has only been run within, and only been used by, the operating system kernel. In a real operating system, the kernel's main function is to provide support for user­level programs. Most such support is accessed via system calls. However, before looking at how to implement system calls, we first walk you through how the kernel gets to run. Remember, that the kernel is just like most programs except that it is the first program that gets to

run. Now that you have some idea on how the kernel boots up, take a look at the comments inside the __start function to get an idea of what it is doing. Try to find the lines where the Interrupt Service Routines (ISRs) are loaded into memory. Because a system call allows flow of information across the userspace­kernelspace boundary, there needs to be relevant code on both sides to handle a system call. We give you one system call, reboot,

which is implemented in the function sys_reboot() in main.c. In GDB, if you put a breakpoint on sys_reboot and run the reboot program, you can print a backtrace in GDB to see how it got there. How a system call occurs in userspace A system call is invoked from userspace by using the syscall instruction which can be seen by disassembling any of the binaries.

The system call numbering has to be agreed upon between userspace and kernelspace. This is taken care of libc that is present in your sources. It does this by using the kernel header file kern/include/kern/syscall.h.

What happens in the kernel on a system call 2.2. Process Support You will also be implementing the subsystem that keeps track of the multiple processes you will have in the future. You must decide what data structures you will need to hold the data pertinent to a process. OS/161 contains some code that should get you started here in kern/include/proc.h. 2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 4/13

It may be helpful to look at kernel include files of your favorite operating system for suggestions, specifically the proc structure. That said, these structures are likely to include much more process (or task) state than you need to correctly complete this assignment. 2.3. User Programs Our System/161 simulator can run normal programs compiled from C. The programs are compiled with a cross­compiler, os161‐gcc. This compiler runs on the host machine and produces MIPS binaries and is

the same compiler used to compile the OS/161 kernel. To create new user programs, you will need to edit the Makefile in bin, sbin, or testbin (depending on where you put your programs) and then create a directory similar to those that already exist. Use an existing program and its Makefile as a template. You are strongly encouraged to run the tests that already exist and write your own as needed.

3. Design Beginning with ASST2 your design documents become an important part of approaching each assignment. A design document should clearly reflect the development of your solution, not merely explain what you

programmed. If you try to code first and design later, or even if you design hastily and rush into coding, you will most certainly end up confused and frustrated. Don't do it! Work with your partner to plan everything you will do. Don't even think about coding until you can precisely explain to each other what

problems you need to solve and how the pieces relate to each other.

Note that it can often be hard to write (or talk) about new software design—you are facing problems that you have not seen before, and therefore even finding terminology to describe your ideas can be difficult. There is no magic solution to this problem, but it gets easier with practice. The important thing is to go ahead and try. Always try to describe your ideas and designs to your partner. In order to reach an understanding, you may have to invent terminology and notation—this is fine, just be sure to document it in your design. If you do this, by the time you have completed your design, you will find that you have the

ability to efficiently discuss problems that you have never seen before. Your design document can be as long as you like. It should include both English definitions and explanations of core functions and interfaces as well as pseudocode and function definitions where appropriate. We suggest that you use a markup language to format your design nicely. Both Markdown  and AsciiDoc  are supported by many Git hosting sites. The contents of your design document should include (but not be limited to):

1. A description of each new piece of functionality you need to add for ASST2. 2. A list and brief description of any new data structures you will have to add to the system. 3. Indications of what, if any, existing code you may model your solution off of. 4. A description of how accesses to shared state will be synchronized, if necessary.

5. A breakdown of who will do what between the partners, and a timeline indicating when assignment tasks will be finished and when testing will take place. 2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 5/13

3.1. Design Considerations Here are some additional questions and thoughts to aid in writing your design document. They are not,

by any means, meant to be a comprehensive list of all the issues you will want to consider. You do not need to explicitly answer or discuss these questions in your executive summary, but you should at least think about them. Your system must allow user programs to receive arguments from the command line. For example, you should be able to run the following program:

char *filename = "/bin/cp"; char *args[4]; pid_t pid; args[0] = "cp"; args[1] = "file1"; args[2] = "file2"; args[3] = NULL; pid = fork(); if (pid == 0) { execv(filename, args);

} The code snippet above loads the executable file /bin/cp, installs it as a new process, and executes it. The new process will then find file1 on the disk and copy it to file2. Passing arguments from one user program, through the kernel, into another user program, is a bit of a

chore. What form does this take in C? This is rather tricky, and there are many ways to be led astray. You will probably find that very detailed pictures and several walkthroughs will be most helpful. This piece of code, in particular, is impossible to write correctly without being carefully designed beforehand 3 . Some other questions to consider: 1. What primitive operations exist to support the transfer of data to and from kernel space? Do you want to implement more on top of these?

2. When implementing exec, how will you determine: a. the stack pointer initial value b. the initial register contents

c. the return value d. whether you can execute the program at all? 3. You will need to bullet­proof the OS/161 kernel from user program errors. There should be nothing a user program can do—and we will try almost everything—to crash the operating system, with the exception of explicitly asking the system to halt.

4. What new data structures will you need to manage multiple processes? 5. What relationships do these new structures have with the rest of the system? 2016. 4. 22. ops­class.org | System Calls

https://www.ops­class.org/asst/2/ 6/13 6. How will you manage file accesses? When the shell invokes the cat command, and the cat command starts to read file1, what will happen if another program also tries to read file1? What would you like to happen? 4. Code Reading To help you get started designing, we have provided the following questions as a guide

for reading through the code. We recommend that you and your partner look over and answer the code reading questions together. You may want to start by reviewing the questions separately, but then meet to talk over your answers and look at the code together. Once you have done this, you should be ready to discuss strategy for designing your code for this assignment. A big part of ASST2 is figuring out what to do, not just how to do it.

4.1. Existing Process Support The key files that are responsible for the loading and running of user­level programs are loadelf.c, runprogram.c, and uio.c, although you may want to add more of your own during this assignment. Understanding these files is the key to getting started with the implementation of multiprogramming. Note that to answer some of the questions, you will have to look in other files. 4.1.1. kern/syscall/loadelf.c This file contains the functions responsible for loading an ELF 4 executable from the file system and

into virtual memory space. Of course, at this point this virtual memory space does not provide what is normally meant by virtual memory—although there is translation between virtual and physical addresses, there is no mechanism for providing more memory than exists physically. You will fix this during ASST3. For now, don't worry about it. 4.1.2. kern/syscall/runprogram.c

This file contains only one function, runprogram, which is responsible for running a program from the kernel menu. It is a good base for writing the execv system call, but only a base—when writing your design doc, you should determine what more is required for execv that runprogram() does not concern

itself with. Additionally, once you have designed your process system, runprogram should be altered to start processes properly within this framework. For example, a program started by runprogram should have the standard file descriptors available while it's running. 4.1.3. kern/lib/uio.c This file contains functions for moving data between kernel and user space. Knowing when and how to cross this boundary is critical to properly implementing user programs, so this is a good file to read very

carefully. You should also examine the code in kern/vm/copyinout.c. 4.1.4. Questions to consider 1. What are the ELF magic numbers? 2. What is the difference between UIO_USERISPACE and UIO_USERSPACE? When should one use UIO_SYSSPACE instead? 2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 7/13 3. Why can the struct uio that is used to read in a segment be allocated on the stack in

load_segment? Or, put another way, where does the memory read actually go? 4. In runprogram why is it important to call vfs_close before going to user mode? 5. What function forces the processor to switch into user mode? Is this function machine dependent? 6. In what files are copyin, copyout, and memmove defined? Why are copyin and copyout necessary?

(As opposed to just using memmove.) 7. What is the purpose of userptr_t? 4.2. kern/arch/mips: Traps and System Calls Exceptions are the key to operating systems. They are the mechanism that enables the operating

system to regain control of execution and therefore do its job. You can think of exceptions as the interface between the processor and the operating system. When the OS boots, it installs an exception handler, usually carefully crafted assembly code, at a specific address in memory. When the processor raises an exception, it invokes this, which sets up a trap frame

4.2.1. locore/trap.c mips_trap is the key function for returning control to the operating system. This is the C function that gets called by the assembly exception handler. enter_new_process is the key function for returning control to user programs. kill_curthread is the function for handling broken user programs. When the processor is in user mode and hits something it can't handle—a bad instruction or our favorite divide­by­zero—it raises an exception. There's no way to recover from this, so the OS needs to kill off the process. Part of

this assignment is writing a useful version of this function. 4.2.2. syscall/syscall.c syscall is the function that delegates the actual work of a system call to the kernel function that implements it. Notice that reboot is the only case currently handled. You will also find a function, enter_forked_process, which is a stub where you will place your code to implement the fork system call.

It should get called from sys_fork. 4.2.3. Questions to consider 1. What is the numerical value of the exception code for a MIPS system call?

2. How many bytes is an instruction in MIPS? Try to answer this by reading syscall carefully, not by looking somewhere else. 3. Why do you probably want to change the implementation of kill_curthread? 4. What would be required to implement a system call that took more than 4 arguments? 4.3. Support Code for User Programs

2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 8/13 As important as the kernel implementation of system calls is the user space code that allows C programs to use them. This code can be found under the userland/ directory at the top of your OS/161 source

tree. Note that you can complete this assignment without modifying user­level code, and you should not break any interface conventions already present—don't swap the order of the argc and argv * arguments to main, for example. However, it is useful to understand how this code works and how it

implements the other side of the system call interface. 4.3.1. userland/lib/crt0/mips/ This is the user program startup code. There's only one file in here, mips‐crt0.S, which contains the

MIPS assembly code that receives control first when a user­level program is started. It calls the user program's main. This is the code that your execv implementation will be interfacing to, so be sure to check what values it expects to appear in what registers and so forth. 4.3.2. userland/lib/libc/ This is the user­level C library. There's obviously a lot of code here. We don't expect you to read it all, although it may be instructive in the long run to do so. For present purposes you need only look at the code that implements the user­level side of system calls, which we detail below.

4.3.3. userland/lib/libc/unix/errno.c This is where the global variable errno is defined. 4.3.4. userland/lib/libc/arch/mips/syscalls‐mips.S This file contains the machine­dependent code necessary for implementing the user­level side of MIPS

system calls. 4.3.5. build/userland/lib/libc/syscalls.S This file is created from syscalls‐mips.S at compile time and is the actual file assembled into the C library. The actual names of the system calls are placed in this file using a script called syscalls/gensyscalls.sh that reads them from the kernel's header files. This avoids having to make a second list of the system calls. In a real system, typically each system call stub is placed in its own

source file, to allow selectively linking them in. OS/161 puts them all together to simplify the build. 4.3.6. Questions to consider 1. What is the purpose of the SYSCALL macro?

2. What is the MIPS instruction that actually triggers a system call? (Answer this by reading the source in this directory, not looking somewhere else.) 3. Now that OS/161 supports 64 bit values, lseek takes and returns a 64 bit offset value. Thus, lseek takes a 32 bit file handle (arg0), a 64 bit offset (arg1), a 32 bit whence (arg3), and needs to return

a 64 bit offset. In void syscall(struct trapframe *tf) where will you find each of the three arguments (in which registers) and how will you return the 64 bit offset? 5. Implementation

Implement system calls and exception handling. 2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 9/13

The full range of system calls that we think you might want over the course of the semester is listed in kern/include/kern/syscall.h. For this assignment you should implement: File system support: open, read, write, lseek, close, dup2, chdir, and __getcwd. Process support: getpid, fork, execv, waitpid, and _exit. It's crucial that your system calls handle all error conditions gracefully without crashing your kernel. You should consult the OS/161 man pages included in the distribution under man/syscall and understand fully the system calls that you must implement. You must return the error codes as described in the man

pages. Additionally, your system calls must return the correct value (in case of success) or error code (in case of failure) as specified in the man pages. The grading scripts rely on the return of appropriate error codes and so adherence to the guidelines is as important as the correctness of your implementation. The file userland/include/unistd.h contains the user­level interface definition of the system calls that

you will be writing for OS/161. This interface is different from that of the kernel functions that you will define to implement these calls. You need to design this interface and put it in kern/include/syscall.h. As you discovered in ASST0, the integer codes for the calls are defined in kern/include/kern/syscall.h. You need to think about a variety of issues associated with implementing system calls. Perhaps the most

obvious one: can two different user­level processes (or user­level threads, if you choose to implement them) find themselves running a system call at the same time? Be sure to argue for or against this, and explain your final decision in your design document. 5.1. Kernel Menu Changes A small but important part of ASST2 is improving the relationship between the kernel menu thread and the thread that it creates that will run your shell (via s) or user programs (via p). Currently the menu

thread does not coordinate with that it forks to go to user space. This means that the kernel menu will immediately return to the top of its loop, redraw the prompt, and begin trying to read terminal input. If the user program that was launched is also trying to read from the terminal (like /bin/shell), it will compete with the kernel menu thread for input. If you find yourself having to type //bbiinn//ttrruuee to

run /bin/true, this what is happening 5 . If the user program is just generating output, you will notice that the kernel prompt is printed and interleaved with its output. Given that this makes it impossible to correctly interact with the shell or capture test output—which test161  depends on—fixing this problem is a part of ASST2. There are multiple approaches that will

work. If you consider this while designing sys_wait and sys_exit, it is possible to reuse that approach without duplicating any code. Alternatively, you can fix this problem before implementing any of your system calls by using some of the synchronization primitives you became familiar with during ASST1.

5.2. File System Support For any given process, the first file descriptors (0, 1, and 2) are considered to be standard input (stdin), standard output (stdout), and standard error (stderr). These file descriptors should start out attached to the console device ("con:"), but your implementation must allow programs to use dup2 to change them

to point elsewhere. 2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 10/13 Although these system calls may seem to be tied to the file system, in fact, they are really about manipulation of file descriptors, or process­specific file system state. A large part of this assignment is

designing and implementing a system to track this state. Some of this information—such as the current working directory—is specific only to the process, but other information—such as file offset—is specific to the process and file descriptor. Don't rush this design. Think carefully about the state you need to maintain, how to organize it, and when and how it has to change.

Note that there is a system call __getcwd and then a library routine getcwd. Once you've written the system call, the library routine should function correctly. 5.3. Process Support Process support for ASST2 divides into the easy (getpid) and the not­so­easy: fork, execv, waitpid and _

exit. These system calls are probably the most difficult part of the assignment, but also the most rewarding. They enable multiprogramming and make OS/161 a usable system. 5.3.1. getpid A PID, or process ID, is a unique number that identifies a process. The implementation of getpid is not terribly challenging, but process ID allocation and reclamation are the important concepts that you must implement. It is not OK for your system to crash because over the lifetime of its execution you've used up all the PIDs. Design your PID system, implement all the tasks associated with PID maintenance, and only then implement getpid. 5.3.2. fork fork is the mechanism for creating new processes. It should make a copy of the invoking process and make sure that the parent and child processes each observe the correct return value (that is, 0 for the child and the newly created PID for the parent). You will want to think carefully through the design of fork

and consider it together with execv to make sure that each system call is performing the correct functionality. fork is also likely to be a chance for you to use one of the synchronization primitives you have implemented previously. 5.3.3. execv execv, although merely a system call, is really the heart of this assignment. It is responsible for taking newly created processes and make them execute something different than what the parent is executing. It must replace the existing address space with a brand new one for the new executable—created by

calling as_create in the current dumbvm system—and then run it. While this is similar to starting a process straight out of the kernel, as runprogram does, it's not quite that simple. Remember that this call is coming out of user space, into the kernel, and then returning back to user space. You must manage the

memory that travels across these boundaries very carefully. Also, notice that runprogram doesn't take an argument vector, but this must of course be handled correctly by execv. 5.3.4. waitpid Although it may seem simple at first, waitpid requires a fair bit of design. Read the specification carefully to understand the semantics, and consider these semantics from the ground up in your design. You may

also wish to consult the UNIX man page, though keep in mind that you are not required to implement all the things UNIX waitpid supports, nor is the UNIX parent/child model of waiting the only valid or viable possibility.

2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 11/13 5.3.5. _exit The implementation of exit is intimately connected to the implementation of waitpid: They are

essentially two halves of the same mechanism. Most of the time, the code for _exit will be simple and the code for waitpid relatively complicated, but it's perfectly viable to design it the other way around as well. If you find both are becoming extremely complicated, it may be a sign that you should rethink your

design. waitpid/_exit is _another chance to use your synchronization primitives. 5.3.6. kill_curthread Feel free to write kill_curthread in as simple a manner as possible. Just keep in mind that essentially nothing about the current thread's user space state can be trusted if it has suffered a fatal exception. It

the user level. 5.4. Error Handling The man pages in the OS/161  distribution contain a description of the error return values that you must return. If there are conditions that can happen that are not listed in the man page, return the most

appropriate error code from kern/include/kern/errno.h. If none seem particularly appropriate, consider adding a new one. If you're adding an error code for a condition for which Unix has a standard error code symbol, use the same symbol if possible. If not, feel free to make up your own, but note that error codes should always begin with E, should not be EOF, etc. Consult Unix man pages to learn about Unix error codes. On Linux systems man errno will do the trick. Note that if you add an error code to kern/include/kern/errno.h, you need to add a corresponding error

message to the file kern/include/kern/errmsg.h. 5.5. kern/proc/proc.c This file contains functions that allow process management. Functions proc_create and proc_destroy are essentially the entry points into process creation and destruction. Keep in mind that you will have to add code to these functions to manage any additional fields that you add to the process structure.

Since a process can contain multiple threads, the functions proc_addthread and proc_remthread are provided to enable tracking. Since we are not going to implement multi­threading, you will not need to modify these functions, but going through the code and understanding how the link between a process

and its threads is established would be beneficial. The functions proc_getas and proc_setas are essentially the getter and setter for the address space that a process has been assigned. You will find an example of how to use proc_setas by looking into the runprogram function. As a final note, address space management is one of the core objectives of ASST3. Finally, In OS/161, there is also a process created for the kernel which tracks all the kernel threads that are created. This process is created by calling proc_bootstrap. Try to track down when and where this

happens. 6. Scheduling 2016. 4. 22. ops­class.org | System Calls

https://www.ops­class.org/asst/2/ 12/13 Although implementing a scheduler is not part of the assignments, the scheduler is a critical components of most operating system. Your OS/161 code also implements a scheduler. A scheduler's primary job is to determine the order in which processes should run. Once

the order is determined, the kernel needs to switch between processes. Or in other words, the kernel needs to perform a context­switch. 7. Writing Your Own Tests Part of ASST2 is learning to test your own code. We provide comprehensive test programs that usually test multiple system calls at once. A consequence of this is that if even one of these system calls is unimplemented, then the whole test will likely fail. You are encouraged to write your own tests which will help you debug specific system calls. You are also encouraged to look at /testbin/badcall and refer to the OS/161 man pages to ensure that you are handling all of the exceptions correctly.

8. Grading ASST2 grading is divided into three parts: 1. File­related system calls 2. Process­related system calls 3. Stability 8.1. File­related System Calls

We will test the following: 1. Does your console work? We will use /testbin/consoletest to determine this. Once you have implemented the write system call, you should have a working console. Note that you should not need to implement open to write to the console. 2. Do your open and close syscalls work? We will test this using /testbin/opentest, /testbin/closetest.

3. Do your read and write syscalls work? We will test these using /testbin/readwritetest and /testbin/fileonlytest. void schedule(void) { /*

* You can write this. If we do nothing, threads will run in * round‐robin fashion. */ }

2016. 4. 22. ops­class.org | System Calls https://www.ops­class.org/asst/2/ 13/13 4. Does your lseek syscall work? We will test this using /testbin/fileonlytest and /testbin/sparsefile.

5. Does your dup2 syscall work? We will test this using /testbin/redirect. Many tests use multiple system calls. For example, although only /testbin/opentest is listed for testing open, almost all of the file­related syscall testing programs use open. Note that you must pass /testbin/consoletest before you can run most of the other tests. This ensures that your kernel menu is not competing with the test output. 8.2. Process­related System Calls

We will test the following: 1. Does your fork syscall work? We will use /testbin/forktest to test this. 2. Does your execv syscall work? We will use /testbin/argtest, /testbin/add, /testbin/factorial and /testbin/bigexec to test this. 3. Do your waitpid/_exit syscalls work? We will use /testbin/forktest, to test this. A lot of other tests internally use fork, waitpid and _exit. You will need to implement waitpid and _exit to pass

these. 4. Does your getpid syscall work? Almost all of the tests described above use getpid in some way. You will need getpid to work to pass these. 8.3. Stability

We will test the stability of your system by running the following tests: 1. /testbin/badcall Tests whether your system call interface can handle invalid arguments. 2. /testbin/crash Tests whether your system can gracefully recover from user programs attempting to perform malicious actions. 3. /testbin/forktest Runs multiple iterations of /testbin/forktest to check for race conditions.

Related Videos Below are some related videos and slides that may help you complete this assignment. 02­28­2014 // Jinghao Shi

03­05­2014 // Guru Prasad 03­26­2014 // Jinghao Shi

02­17­2015 // Jinghao Shi 02­24­2015 // Jinghao Shi 03­03­2015 // Jinghao Shi 02­23­2016 // Ali Ben Ali // Slides  03­01­2016 // Ali Ben Ali // Slides  03­08­2016 // Ali Ben Ali // Slides 