Linux I/O Models
2015-09-29 11:55
1046 查看
http://www.ibm.com/developerworks/library/l-async/
The most common input/output (I/O) model used in Linux® is synchronous I/O. After a request is made in this model, the application blocks until the request is satisfied. This is a great paradigm because the calling applicationrequires no central processing unit (CPU) while it awaits the completion of the I/O request. But in some cases there's a need to overlap an I/O request with other processing. The Portable Operating System Interface (POSIX) asynchronous I/O (AIO) application
program interface (API) provides this capability. In this article, get an overview of the API and see how to use it.
Introduction to AIO
Linux asynchronous I/O is a relatively recent addition to the Linux kernel. It's a standard feature of the 2.6 kernel, but you can find patches for 2.4. The basic idea behind AIO is to allow a process to initiate a number of I/O operations without having toblock or wait for any to complete. At some later time, or after being notified of I/O completion, the process can retrieve the results of the I/O.
I/O models
Before digging into the AIO API, let's explore the different I/O models that are available under Linux. This isn't intended as an exhaustive review, but rather aims to cover the most common models to illustrate their differences from asynchronous I/O. Figure1 shows synchronous and asynchronous models, as well as blocking and non-blocking models.
Figure 1. Simplified matrix of basic Linux I/O models
Each of these I/O models has usage patterns that are advantageous for particular applications. This section briefly explores each one.
Synchronous blocking I/O
I/O-bound versus CPU-bound processes
A process that is I/O bound is one that performs more I/O than processing. A CPU-bound process does more processing than I/O. The Linux 2.6 scheduler actually favors I/O-bound processes because they commonly initiate an I/O and then block, which means otherwork can be efficiently interlaced between them.
One of the most common models is the synchronous blocking I/O model. In this model, the user-space application performs a system call that results in the application blocking. This means that the application blocks until the system call is complete (data transferred
or error). The calling application is in a state where it consumes no CPU and simply awaits the response, so it is efficient from a processing perspective.
Figure 2 illustrates the traditional blocking I/O model, which is also the most common model used in applications today. Its behaviors are well understood, and its usage is efficient for typical applications. When the
readsystem
call is invoked, the application blocks and the context switches to the kernel. The read is then initiated, and when the response returns (from the device from which you're reading), the data is moved to the user-space buffer. Then the application is unblocked
(and the
readcall
returns).
Figure 2. Typical flow of the synchronous blocking I/O model
From the application's perspective, the
readcall
spans a long duration. But, in fact, the application is actually blocked while the read is multiplexed with other work in the kernel.
Synchronous non-blocking I/O
A less efficient variant of synchronous blocking is synchronous non-blocking I/O. In this model, a device is opened as non-blocking. This means that instead of completing an I/O immediately, a readmay
return an error code indicating that the command could not be immediately satisfied (
EAGAINor
EWOULDBLOCK),
as shown in Figure 3.
Figure 3. Typical flow of the synchronous non-blocking I/O model
The implication of non-blocking is that an I/O command may not be satisfied immediately, requiring that the application make numerous calls to await completion. This can be extremely inefficient because in many cases the application must busy-wait until the
data is available or attempt to do other work while the command is performed in the kernel. As also shown in Figure 3, this method can introduce latency in the I/O because any gap between the data becoming available in the kernel and the user calling
readto
return it can reduce the overall data throughput.
Asynchronous blocking I/O
Another blocking paradigm is non-blocking I/O with blocking notifications. In this model, non-blocking I/O is configured, and then the blockingselectsystem
call is used to determine when there's any activity for an I/O descriptor. What makes the
selectcall
interesting is that it can be used to provide notification for not just one descriptor, but many. For each descriptor, you can request notification of the descriptor's ability to write data, availability of read data, and also whether an error has occurred.
Figure 4. Typical flow of the asynchronous blocking I/O model (select)
The primary issue with the
selectcall
is that it's not very efficient. While it's a convenient model for asynchronous notification, its use for high-performance I/O is not advised.
Asynchronous non-blocking I/O (AIO)
Finally, the asynchronous non-blocking I/O model is one of overlapping processing with I/O. The read request returns immediately, indicating that the readwas
successfully initiated. The application can then perform other processing while the background read operation completes. When the
readresponse
arrives, a signal or a thread-based callback can be generated to complete the I/O transaction.
Figure 5. Typical flow of the asynchronous non-blocking I/O model
The ability to overlap computation and I/O processing in a single process for potentially multiple I/O requests exploits the gap between processing speed and I/O speed. While one or more slow I/O requests are pending, the CPU can perform other tasks or, more
commonly, operate on already completed I/Os while other I/Os are initiated.
The next section examines this model further, explores the API, and then demonstrates a number of the commands.
Back
to top
Motivation for asynchronous I/O
From the previous taxonomy of I/O models, you can see the motivation for AIO. The blocking models require the initiating application to block when the I/O has started. This means that it isn't possible to overlap processing and I/O at the same time. The synchronousnon-blocking model allows overlap of processing and I/O, but it requires that the application check the status of the I/O on a recurring basis. This leaves asynchronous non-blocking I/O, which permits overlap of processing and I/O, including notification of
I/O completion.
The functionality provided by the
selectfunction
(asynchronous blocking I/O) is similar to AIO, except that it still blocks. However, it blocks on notifications instead of the I/O call.
Back
to top
Introduction to AIO for Linux
This section explores the asynchronous I/O model for Linux to help you understand how to apply it in your applications.In a traditional I/O model, there is an I/O channel that is identified by a unique handle. In UNIX®, these are file descriptors (which are the same for files, pipes, sockets, and so on). In blocking I/O, you initiate a transfer and the system call returns when
it's complete or an error has occurred.
AIO for Linux
AIO first entered the Linux kernel in 2.5 and is now a standard feature of 2.6 production kernels.In asynchronous non-blocking I/O, you have the ability to initiate multiple transfers at the same time. This requires a unique context for each transfer so you can identify it when it completes. In AIO, this is an
aiocb(AIO
I/O Control Block) structure. This structure contains all of the information about a transfer, including a user buffer for data. When notification for an I/O occurs (called a completion), the
aiocbstructure
is provided to uniquely identify the completed I/O. The API demonstration shows how to do this.
Back
to top
AIO API
The AIO interface API is quite simple, but it provides the necessary functions for data transfer with a couple of different notification models. Table 1 shows the AIO interface functions, which are further explained later in this section.Table 1. AIO interface APIs
API function | Description |
---|---|
aio_read | Request an asynchronous read operation |
aio_error | Check the status of an asynchronous request |
aio_return | Get the return status of a completed asynchronous request |
aio_write | Request an asynchronous operation |
aio_suspend | Suspend the calling process until one or more asynchronous requests have completed (or failed) |
aio_cancel | Cancel an asynchronous I/O request |
lio_listio | Initiate a list of I/O operations |
aiocbstructure
for initiating or checking. This structure has a number of elements, but Listing 1 shows only the ones that you'll need to (or can) use.
Listing 1. The aiocb structure showing the relevant fields
struct aiocb { int aio_fildes; // File Descriptor int aio_lio_opcode; // Valid only for lio_listio (r/w/nop) volatile void *aio_buf; // Data Buffer size_t aio_nbytes; // Number of Bytes in Data Buffer struct sigevent aio_sigevent; // Notification Structure /* Internal fields */ ... };
The
sigeventstructure
tells AIO what to do when the I/O completes. You'll explore this structure in the AIO demonstration. Now I'll show you how the individual API functions for AIO work and how you can use them.
aio_read
The aio_readfunction
requests an asynchronous read operation for a valid file descriptor. The file descriptor can represent a file, a socket, or even a pipe. The
aio_readfunction
has the following prototype:
int aio_read( struct aiocb *aiocbp );
The
aio_readfunction
returns immediately after the request has been queued. The return value is zero on success or -1 on error, where
errnois
defined.
To perform a read, the application must initialize the
aiocbstructure.
The following short example illustrates filling in the
aiocbrequest
structure and using
aio_readto
perform an asynchronous read request (ignore notification for now). It also shows use of the
aio_errorfunction,
but I'll explain that later.
Listing 2. Sample code for an asynchronous read with aio_read
#include <aio.h> ... int fd, ret; struct aiocb my_aiocb; fd = open( "file.txt", O_RDONLY ); if (fd < 0) perror("open"); /* Zero out the aiocb structure (recommended) */ bzero( (char *)&my_aiocb, sizeof(struct aiocb) ); /* Allocate a data buffer for the aiocb request */ my_aiocb.aio_buf = malloc(BUFSIZE+1); if (!my_aiocb.aio_buf) perror("malloc"); /* Initialize the necessary fields in the aiocb */ my_aiocb.aio_fildes = fd; my_aiocb.aio_nbytes = BUFSIZE; my_aiocb.aio_offset = 0; ret = aio_read( &my_aiocb ); if (ret < 0) perror("aio_read"); while ( aio_error( &my_aiocb ) == EINPROGRESS ) ; if ((ret = aio_return( &my_iocb )) > 0) { /* got ret bytes on the read */ } else { /* read failed, consult errno */ }
In Listing 2, after the file from which you're reading data is opened, you zero out your
aiocbstructure,
and then allocate a data buffer. The reference to the data buffer is placed into
aio_buf.
Subsequently, you initialize the size of the buffer into
aio_nbytes.
The
aio_offsetis
set to zero (the first offset in the file). You set the file descriptor from which you're reading into
aio_fildes.
After these fields are set, you call
aio_readto
request the read. You can then make a call to
aio_errorto
determine the status of the
aio_read.
As long as the status is
EINPROGRESS,
you busy-wait until the status changes. At this point, your request has either succeeded or failed.
Building with the AIO interface
You can find the function prototypes and other necessary symbolics in the aio.hheader
file. When building an application that uses this interface, you must use the POSIX real-time extensions library (
librt).
Note the similarities to reading from the file with the standard library functions. In addition to the asynchronous nature of
aio_read,
another difference is setting the offset for the read. In a typical
readcall,
the offset is maintained for you in the file descriptor context. For each read, the offset is updated so that subsequent reads address the next block of data. This isn't possible with asynchronous I/O because you can perform many read requests simultaneously,
so you must specify the offset for each particular read request.
aio_error
The aio_errorfunction
is used to determine the status of a request. Its prototype is:
int aio_error( struct aiocb *aiocbp );
This function can return the following:
EINPROGRESS,
indicating the request has not yet completed
ECANCELLED,
indicating the request was cancelled by the application
-1, indicating
that an error occurred for which you can consult
errno
aio_return
Another difference between asynchronous I/O and standard blocking I/O is that you don't have immediate access to the return status of your function because you're not blocking on the readcall.
In a standard
readcall,
the return status is provided upon return of the function. With asynchronous I/O, you use the
aio_returnfunction.
This function has the following prototype:
ssize_t aio_return( struct aiocb *aiocbp );
This function is called only after the
aio_errorcall
has determined that your request has completed (either successfully or in error). The return value of
aio_returnis
identical to that of the
reador
writesystem
call in a synchronous context (number of bytes transferred or
-1for
error).
aio_write
The aio_writefunction
is used to request an asynchronous write. Its function prototype is:
int aio_write( struct aiocb *aiocbp );
The
aio_writefunction
returns immediately, indicating that the request has been enqueued (with a return of
0on
success and
-1on
failure, with
errnoproperly
set).
This is similar to the
readsystem
call, but one behavior difference is worth noting. Recall that the offset to be used is important with the
readcall.
However, with
write,
the offset is important only if used in a file context where the
O_APPENDoption
is not set. If
O_APPENDis
set, then the offset is ignored and the data is appended to the end of the file. Otherwise, the
aio_offsetfield
determines the offset at which the data is written to the file.
aio_suspend
You can use the aio_suspendfunction
to suspend (or block) the calling process until an asynchronous I/O request has completed, a signal is raised, or an optional timeout occurs. The caller provides a list of
aiocbreferences
for which the completion of at least one will cause
aio_suspendto
return. The function prototype for
aio_suspendis:
int aio_suspend( const struct aiocb *const cblist[], int n, const struct timespec *timeout );
Using
aio_suspendis
quite simple. A list of
aiocbreferences
is provided. If any of them complete, the call returns with
0.
Otherwise,
-1is
returned, indicating an error occurred. See Listing 3.
Listing 3. Using the aio_suspend function to block on asynchronous I/Os
struct aioct *cblist[MAX_LIST] /* Clear the list. */ bzero( (char *)cblist, sizeof(cblist) ); /* Load one or more references into the list */ cblist[0] = &my_aiocb; ret = aio_read( &my_aiocb ); ret = aio_suspend( cblist, MAX_LIST, NULL );
Note that the second argument of
aio_suspendis
the number of elements in
cblist,
not the number of
aiocbreferences.
Any
NULLelement
in the
cblistis
ignored by
aio_suspend.
If a timeout is provided to
aio_suspendand
the timeout occurs, then
-1is
returned and
errnocontains
EAGAIN.
aio_cancel
The aio_cancelfunction
allows you to cancel one or all outstanding I/O requests for a given file descriptor. Its prototype is:
int aio_cancel( int fd, struct aiocb *aiocbp );
To cancel a single request, provide the file descriptor and the
aiocbreference.
If the request is successfully cancelled, the function returns
AIO_CANCELED.
If the request completes, the function returns
AIO_NOTCANCELED.
To cancel all requests for a given file descriptor, provide that file descriptor and a
NULLreference
for
aiocbp.
The function returns
AIO_CANCELEDif
all requests are canceled,
AIO_NOT_CANCELEDif
at least one request couldn't be canceled, and
AIO_ALLDONEif
none of the requests could be canceled. You can then evaluate each individual AIO request using
aio_error.
If the request was canceled,
aio_errorreturns
-1,
and
errnois
set to
ECANCELED.
lio_listio
Finally, AIO provides a way to initiate multiple transfers at the same time using the lio_listioAPI
function. This function is important because it means you can start lots of I/Os in the context of a single system call (meaning one kernel context switch). This is great from a performance perspective, so it's worth exploring. The
lio_listioAPI
function has the following prototype:
int lio_listio( int mode, struct aiocb *list[], int nent, struct sigevent *sig );
The
modeargument
can be
LIO_WAITor
LIO_NOWAIT.
LIO_WAITblocks
the call until all I/O has completed.
LIO_NOWAITreturns
after the operations have been queued. The
listis
a list of
aiocbreferences,
with the maximum number of elements defined by
nent.
Note that elements of
listmay
be
NULL,
which
lio_listioignores.
The
sigeventreference
defines the method for signal notification when all I/O is complete.
The request for
lio_listiois
slightly different than the typical
reador
writerequest
in that the operation must be specified. This is illustrated in Listing 4.
Listing 4. Using the lio_listio function to initiate a list of requests
struct aiocb aiocb1, aiocb2; struct aiocb *list[MAX_LIST]; ... /* Prepare the first aiocb */ aiocb1.aio_fildes = fd; aiocb1.aio_buf = malloc( BUFSIZE+1 ); aiocb1.aio_nbytes = BUFSIZE; aiocb1.aio_offset = next_offset; aiocb1.aio_lio_opcode = LIO_READ; ... bzero( (char *)list, sizeof(list) ); list[0] = &aiocb1; list[1] = &aiocb2; ret = lio_listio( LIO_WAIT, list, MAX_LIST, NULL );
The read operation is noted in the
aio_lio_opcodefield
with
LIO_READ.
For a write operation,
LIO_WRITEis
used, but
LIO_NOPis
also valid for no operation.
Back
to top
AIO notifications
Now that you've seen the AIO functions that are available, this section digs into the methods that you can use for asynchronous notification. I'll explore asynchronous notification through signals and function callbacks.
Asynchronous notification with signals
The use of signals for interprocess communication (IPC) is a traditional mechanism in UNIX and is also supported by AIO. In this paradigm, the application defines a signal handler that is invoked when a specified signal occurs. The application then specifiesthat an asynchronous request will raise a signal when the request has completed. As part of the signal context, the particular
aiocbrequest
is provided to keep track of multiple potentially outstanding requests. Listing 5 demonstrates this notification method.
Listing 5. Using signals as notification for AIO requests
void setup_io( ... ) { int fd; struct sigaction sig_act; struct aiocb my_aiocb; ... /* Set up the signal handler */ sigemptyset(&sig_act.sa_mask); sig_act.sa_flags = SA_SIGINFO; sig_act.sa_sigaction = aio_completion_handler; /* Set up the AIO request */ bzero( (char *)&my_aiocb, sizeof(struct aiocb) ); my_aiocb.aio_fildes = fd; my_aiocb.aio_buf = malloc(BUF_SIZE+1); my_aiocb.aio_nbytes = BUF_SIZE; my_aiocb.aio_offset = next_offset; /* Link the AIO request with the Signal Handler */ my_aiocb.aio_sigevent.sigev_notify = SIGEV_SIGNAL; my_aiocb.aio_sigevent.sigev_signo = SIGIO; my_aiocb.aio_sigevent.sigev_value.sival_ptr = &my_aiocb; /* Map the Signal to the Signal Handler */ ret = sigaction( SIGIO, &sig_act, NULL ); ... ret = aio_read( &my_aiocb ); } void aio_completion_handler( int signo, siginfo_t *info, void *context ) { struct aiocb *req; /* Ensure it's our signal */ if (info->si_signo == SIGIO) { req = (struct aiocb *)info->si_value.sival_ptr; /* Did the request complete? */ if (aio_error( req ) == 0) { /* Request completed successfully, get the return status */ ret = aio_return( req ); } } return; }
In Listing 5, you set up your signal handler to catch the
SIGIOsignal
in the
aio_completion_handlerfunction.
You then initialize the
aio_sigeventstructure
to raise
SIGIOfor
notification (which is specified via the
SIGEV_SIGNALdefinition
in
sigev_notify).
When your read completes, your signal handler extracts the particular
aiocbfrom
the signal's
si_valuestructure
and checks the error status and return status to determine I/O completion.
For performance, the completion handler is an ideal spot to continue the I/O by requesting the next asynchronous transfer. In this way, when completion of one transfer has completed, you immediately start the next.
Asynchronous notification with callbacks
An alternative notification mechanism is the system callback. Instead of raising a signal for notification, this mechanism calls a function in user-space for notification. You initialize the aiocbreference
into the
sigeventstructure
to uniquely identify the particular request being completed; see Listing 6.
Listing 6. Using thread callback notification for AIO requests
void setup_io( ... ) { int fd; struct aiocb my_aiocb; ... /* Set up the AIO request */ bzero( (char *)&my_aiocb, sizeof(struct aiocb) ); my_aiocb.aio_fildes = fd; my_aiocb.aio_buf = malloc(BUF_SIZE+1); my_aiocb.aio_nbytes = BUF_SIZE; my_aiocb.aio_offset = next_offset; /* Link the AIO request with a thread callback */ my_aiocb.aio_sigevent.sigev_notify = SIGEV_THREAD; my_aiocb.aio_sigevent.notify_function = aio_completion_handler; my_aiocb.aio_sigevent.notify_attributes = NULL; my_aiocb.aio_sigevent.sigev_value.sival_ptr = &my_aiocb; ... ret = aio_read( &my_aiocb ); } void aio_completion_handler( sigval_t sigval ) { struct aiocb *req; req = (struct aiocb *)sigval.sival_ptr; /* Did the request complete? */ if (aio_error( req ) == 0) { /* Request completed successfully, get the return status */ ret = aio_return( req ); } return; }
In Listing 6, after creating your
aiocbrequest,
you request a thread callback using
SIGEV_THREADfor
the notification method. You then specify the particular notification handler and load the context to be passed to the handler (in this case, a reference to the
aiocbrequest
itself). In the handler, you simply cast the incoming
sigvalpointer
and use the AIO functions to validate the completion of the request.
Back
to top
System tuning for AIO
The proc file system contains two virtual files that can be tuned for asynchronous I/O performance:The /proc/sys/fs/aio-nr file provides the current number of system-wide asynchronous I/O requests.
The /proc/sys/fs/aio-max-nr file is the maximum number of allowable concurrent requests. The maximum is commonly 64KB, which is adequate for most applications.
Back
to top
Summary
Using asynchronous I/O can help you build faster and more efficient I/O applications. If your application can overlap processing and I/O, then AIO can help you build an application that more efficiently uses the CPU resources available to you. While this I/Omodel differs from the traditional blocking patterns found in most Linux applications, the asynchronous notification model is conceptually simple and can simplify your design.
相关文章推荐
- Centos学习笔记 by 小陌_0226
- Linux常见目录使用区别
- linux sleep函数
- Linux学习笔记之 Btrfs文件系统简介及使用
- drm 驱动是如何创建 fb device 的
- Linux 性能监控、测试、优化工具
- Linux下的一些快捷键
- Linux优化之IO子系统监控与调优
- 【马哥linux学员学习笔记】MySQL多实例详解
- Linux Pmap 命令:查看进程用了多少内存
- linux查看占用内存/cpu最高的进程情况
- CentOS 7系统添加启动项
- 每天进步一点点——设置Sysctl.conf用以提高Linux的性能(最完整的sysctl.conf优化方案)
- 阿里云主机centos设置虚拟内存(交换分区)
- linux ps命令详解
- Linux设备驱动——字符设备驱动
- LINUX CETOS7
- linux磁盘管理
- Linux mysql 工具常用命令
- Linux文件I/O的lseek,fcntl和ioctl函数