1 / 30

I/O

I/O. CS 416: Operating Systems Design, Spring 2001 Department of Computer Science Rutgers University http://remus.rutgers.edu/cs416/F01/. I/O Devices. So far we have talked about how to abstract and manage CPU and memory

camila
Download Presentation

I/O

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. I/O CS 416: Operating Systems Design, Spring 2001 Department of Computer ScienceRutgers University http://remus.rutgers.edu/cs416/F01/

  2. I/O Devices • So far we have talked about how to abstract and manage CPU and memory • Computation “inside” computer is useful only if some results are communicated “outside” of the computer • I/O devices are the computer’s interface to the outside world (I/O  Input/Output) • Example devices: display, keyboard, mouse, speakers, network interface, and disk CS 416: Operating Systems

  3. Disk Basic Computer Structure CPU Memory Memory Bus (System Bus) Bridge I/O Bus NIC CS 416: Operating Systems

  4. OS: Abstractions and Access Methods • OS must virtualizes a wide range of devices into a few simple abstractions: • Storage • Hard drives, Tapes, CDROM • Networking • Ethernet, radio, serial line • Multimedia • DVD, Camera, microphones • Operating system should provide consistent methods to access the abstractions • Otherwise, programming is too hard CS 416: Operating Systems

  5. User/OS method interface • The same interface is used to access devices (like disks and network lines) and more abstract resources like files • 4 main methods: • open() • close() • read() • write() • Semantics depend on the type of the device (block, char, net) • These methods are system calls because they are the methods the OS provides to all processes. • A Java program can’t access these methods directly, they are used by the JVM. C and C++ programs can call these directly. CS 416: Operating Systems

  6. Unix I/O Methods • fileHandle = open(pathName, flags, mode) • a file handle is a small integer, valid only within a single process, to operate on the device or file • pathname: a name in the file system. In unix, devices are put under /dev. E.g. /dev/ttya is the first serial port, /dev/sda the first SCSI drive • flags: blocking or non-blocking … • mode: read only, read/write, append … • errorCode = close(fileHandle) • Kernel will free the data structures associated with the device CS 416: Operating Systems

  7. Unix I/O Methods • byteCount = read(fileHandle, byte [] buf, count) • read at most count bytes from the device and put them in the byte buffer buf. Bytes placed from 0th byte. • Kernel can give the process less bytes, user process must check the byteCount to see how many were actually returned. • A negative byteCount signals an error (value is the error type) • byteCount = write(fileHandle, byte [] buf, count) • write at most count bytes from the buffer buf • actual number written returned in byteCount • a Negative byteCount signals an error CS 416: Operating Systems

  8. Unix I/O Example • What’s the correct way to write a 1000 bytes? • calling: • ignoreMe = write(fileH, buffer, 1000); • works most of the time. What happens if: • can’t accept 1000 bytes right now? • disk is full? • someone just deleted the file? • Let’s work this out on the board CS 416: Operating Systems

  9. I/O semantics • From this basic interface, three different dimension to how I/O is processed: • blocking vs. non-blocking • buffered vs. unbuffered • synchronous vs. asynchronous • The O/S tries to support as many of these dimensions as possible for each device • The semantics are specified during the open() system call CS 416: Operating Systems

  10. Blocking vs. Non-Blocking I/O • Blocking –process is suspended until all bytes in the count field are read or written • E.g., for a network device, if the user wrote 1000 bytes, then the operating system would write the bytes to the device one at a time until the write() completed. • + Easy to use and understand • - if the device just can’t perform the operation (e.g. you unplug the cable), what to do? Give up an return the successful number of bytes. • Nonblocking – the OS only reads or writes as many bytes as is possible without suspending the process • + Returns quickly • - more work for the programmer (but really for robust programs)? CS 416: Operating Systems

  11. Buffered vs. Unbuffered I/O • Sometime we want the ease of programming of blocked I/O without the long waits if the buffers on the device are small. • buffered I/O allows the kernel to make a copy of the data • write() side: allows the process to write() many bytes and continue processing • read() side: As device signals data is ready, kernel places data in the buffer. When the process calls read(), the kernel just copies the buffer. • Why not use buffered I/O? • - Extra copy overhead • - Delays sending data CS 416: Operating Systems

  12. Synchronous vs. Asynchronous I/O • Synchronous I/O: the user process does not run at the same time the I/O does --- it is suspended during I/O • So far, all the methods presented have been synchronous. • Asynchronous I/O: The read() or write() call returns a small object instead of a count. • separate set of methods in unix: aio_read(), aio_write() • The user can call methods on the returned object to check “how much” of the I/O has completed • The user can also allow a signal handler to run when the the I/O has completed. CS 416: Operating Systems

  13. Handling Multiple I/O streams • If we use blocking I/O, how do we handle > 1 device at a time? • Example : a small program that reads from serial line and outputs to a tape and is also reading from the tape and outputting to a serial line • Structure the code like this? • while (TRUE) { • read(tape); // block until tape is ready • write(serial line);// send data to serial line • read(serial line); • write(tape); • } • Could use non-blocking I/O, but huge waste of cycles if data is not ready. CS 416: Operating Systems

  14. Solution: select system call • totalFds = select(nfds, readfds, writefds, errorfds, timeout); • nfds: the range (0.. nfds) of file descriptors to check • readfds: bit map of fileHandles. user sets the bit X to ask the kernel to check if fileHandle X ready for reading. Kernel returns a 1 if data can be read on the fileHandle. • writefds: bit map if fileHandle Y for writing. Operates same as read bitmap. • errorfds: bit map to check for errors. • timeout: how long to wait before un-suspending the process • totalFds = number of set bits, negative number is an error CS 416: Operating Systems

  15. Three Device Types • Most operating system have three device types: • Character devices • Used for serial-line types of devices (e.g. USB port) • Network devices • Used for network interfaces (E.g. Ethernet card) • Block devices: • Used for mass-storage (E.g. disks and CDROM) • What you can expect from the read/write methods changes with each device type CS 416: Operating Systems

  16. Character Devices • Device is represented by the OS as an ordered stream of bytes • bytes sent out the device by the write() system call • bytes read from the device by the read() system call • Byte stream has no “start”, just open and start/reading writing • The user has no control of the read/write ratio, the sender process might write() 1 time of 1000 bytes, and the receiver may have to call 1000 read calls each receiving 1 byte. CS 416: Operating Systems

  17. Network Devices • Like I/O devices, but each write() call either sends the entire block (packet), up to some maximum fixed size, or none. • On the receiver, the read() call returns all the bytes in the block, or none. CS 416: Operating Systems

  18. Block Devices • OS presents device as a large array of blocks • Each block has a fixed size (1KB - 8KB is typical) • User can read/write only in fixed sized blocks • Unlike other devices, block devices support random access • We can read or write anywhere in the device without having to ‘read all the bytes first’ • But how to specify the nth block with current interface? CS 416: Operating Systems

  19. The file pointer • O/S adds a concept call the file pointer • A file pointer is associated with each open file, if the device is a block device • the next read or write operates at the position in the device pointed to by the file pointer • The file pointer is measured in bytes, not blocks CS 416: Operating Systems

  20. Seeking in a device • To set the file pointer: • absoluteOffset = lseek(fileHandle, offset, whence); • whence specifies if the offset is absolute, from byte 0, or relative to the current file pointer position • the absolute offset is returned; negative numbers signal error codes • For devices, the offset should be a integral number of bytes relative to the size of a block. • How could you tell what the current position of the file pointer is without changing it? CS 416: Operating Systems

  21. Block Device Example • You want to read the 10th block of a disk • each disk block is 2048 bytes long • fh = open(/dev/sda, , , ); • pos = lseek(fh, size*count, ); • if (pos < 0 ) error; • bytesRead = read(fh, buf, blockSize); • if (bytesRead < 0) error; • … CS 416: Operating Systems

  22. Getting and setting device specific information • Unix has an IO ConTroL system call: • ErrorCode = ioctl(fileHandle, int request, object); • request is a numeric command to the device • can also pass an optional, arbitrary object to a device • the meaning of the command and the type of the object are device specific CS 416: Operating Systems

  23. Programmed I/O vs. DMA • Programmed I/O is ok for sending commands, receiving status, and communication of a small amount of data • Inefficient for large amount of data • Keeps CPU busy during the transfer • Programmed I/O  memory operations  slow • Direct Memory Access • Device read/write directly from/to memory • Memory  device typically initiated from CPU • Device  memory can be initiated by either the device or the CPU CS 416: Operating Systems

  24. Disk Disk Disk Programmed I/O vs. DMA CPU Memory CPU Memory CPU Memory Interconnect Interconnect Interconnect Programmed I/O DMA DMA Device  Memory Problems? CS 416: Operating Systems

  25. Direct Memory Access • Used to avoid programmed I/O for large data movement • Requires DMA controller • Bypasses CPU to transfer data directly between I/O device and memory CS 416: Operating Systems

  26. Six step process to perform DMA transfer CS 416: Operating Systems

  27. Life Cycle of an I/O Request CS 416: Operating Systems

  28. Performance • I/O a major factor in system performance • Demands CPU to execute device driver, kernel I/O code • Context switches due to interrupts • Data copying • Network traffic especially stressful CS 416: Operating Systems

  29. Intercomputer communications CS 416: Operating Systems

  30. Improving Performance • Reduce number of context switches • Reduce data copying • Reduce interrupts by using large transfers, smart controllers, polling • Use DMA • Balance CPU, memory, bus, and I/O performance for highest throughput CS 416: Operating Systems

More Related