Home | Projects | Notes > Operating Systems > I/O Management
I/O Management
Techniques for Performing I/O
Programmed I/O
In early systems, the processor issued commands, on behalf of a process, to an I/O module; that process then busy waits for the operation to be completed before proceeding.
Very inefficient use of the processor to control I/O as well as general computation in processes, because I/O devices are comparatively slower than main memory and processor.
Interrupt-Driven I/O
The processor issues the I/O command on behalf of the process. There are two possibilities:
If the I/O instruction from the process is non-blocking, then the processor continues to execute instructions from the process that issued the I/O command.
If the I/O instruction is blocking, then the next instruction that the processor executes is from the OS, which will put the current process in a blocked state and schedule another process.
No busy waiting, but still slow because interrupt is triggered for every little piece of data being transferred.
DMA (Direct Memory Access)
A DMA module controls the exchange of data between main memory and an I/O module. The processor sends a request for the transfer of a block of data to the DMA module, and is interrupted only after the entire block has been transferred.
No busy waiting. Much faster than other previous techniques.
e.g., In virtual memory paging, DMA can be involved in bring in the pages (block transfer) from disk that are not in memory.
[!] Note: In general, block size is either equal to the frame size or multiple of frame size.
Direct Memory Access (DMA)
Above diagram indicates, in terms, the DMA logic. The DMA unit is capable of mimicking the processor and, indeed, of taking over control of the system bus just like a processor. It needs to do this to transfer data to and from memory over the system bus.
The DMA technique works as follows. When the processor wishes to read or write a block of data, it issues a command to the DMA module by sending to the DMA module the following information:
Whether a read or write is requested, using the read or write control line between the processor and the DMA module
The address of the I/O device involved, communicated on the data lines
The starting location in memory to read from or write to, communicated on the data lines and stored by the DMA module in its address register
The number of words to be read or written, again communicated via the data lines stored in the data count register.
The processor then continues with other work. It has delegated this I/O operation to the DMA module.
When the transfer is complete, the DMA module sends an interrupt signal to the processor. Thus, the processor is involved only at the beginning and end of the transfer.
Evolution of the I/O Function
Processor directly controls a peripheral device or directly commands an I/O module attached to that device. i.e., Programmed I/O.
Processor sends instructions to devices as in (1), but interrupts are now used to eliminate busy waiting for those instructions to complete.
The I/O module/device is given direct access to the memory via DMA. It can now move an entire block of data without involving the processor, once started.
The I/O module is enhanced by giving it a processor with a specialized instruction set for the device it manages. Now, a single command can initiate a sequence of instructions.
The I/O module is enhanced with its own local memory and ability to implement customizable driver software to control a device. This is a specialized processor we mentioned in the context of multiprocessing architectures.
전반적으로, OS의 I/O에의 개입이 점점 줄어드는 방향으로 진화해왔다고 보면 된다. 최종적으로는, OS가 다른 process들을 처리하는 중에 I/O를 전담하는 장치가 I/O 작업을 수행하는 방식으로 개선되었다.
Operating System Design Goals
Two objectives in designing the I/O facility: Efficiency and generality.
In other words, the OS can use techniques to speed up I/O (read/write) and it should have general purpose behavior for classes of devices.
e.g., using available memory frames to buffer I/O speeds up a process' ability to use some of the data while waiting for the rest. (i.e., I/O buffering)
e.g.,, Linux categorizes devices as block or stream oriented. Block devices read/write entire blocks at a time, while stream oriented handles one byte at a time, regardless of the actual device type. (Everything in Linux is treated as a file. Reading/writing to/from an I/O device is no different than doing so to/from a file. Uniform operations!)
I/O Buffering
It is sometimes convenient to perform input transfers in advance (anticipation) of requests being needed and to perform output transfers some time after the data is ready.
Buffering data into memory frames is a technique that allows this.
Buffering helps smooth out delays associated with bursts of I/O in a process.
I/O Buffering Techniques
Single Buffering
When a process issues an I/O request, the OS transfers data into the OS. I/O is transferred in/out of that buffer, then the buffer is exchanged into the process' frame(s) of memory.
Meanwhile, the OS can signal another buffer to begin transfer (looking ahead, prefetching, etc.).
The user process can be working with one block of data while the next block is being read for (or older block is being written out).
At any point in time, the user process is working with a single buffer and the OS has a single buffer on hand to dedicate to that process' next transaction.
Double Buffering
The OS reserves two system buffers for transacting I/O on behalf of a user process. When one buffer fills and is being transferred to the user process' memory space, the 2nd buffer can be used to continue working with the I/O devices.
This allows the OS to continue filling a buffer while the user buffer is being transferred over.
Circular Buffering
This is basically an extension from 2 to N buffers.
The OS reserves N system buffers for transacting I/O on behalf of a user process. When one buffer fills and is being transferred to the user process' memory space, the next buffer in line can be used to continue working with the I/O devices.
This works well for systems with I/O devices of varying speeds. If the OS has a large reserve of memory to work with, it can allocate more buffers to each process.
This allows the OS to pre-page the anticipated pages (based on the Principle of Locality) as much as possible.
References
Stallings, W. (2018). Operating Systems: Internals and Design Principles (9th ed.). Pearson Education, Inc.