[ViewVC] Annotation of: cvs/Linux-AIO/README

NAME
    Linux::AIO - linux-specific aio implemented using clone

SYNOPSIS
     use Linux::AIO;

DESCRIPTION
    This module implements asynchronous I/O using the means available to
    Linux - clone. It does not hook into the POSIX aio_* functions because
    Linux does not yet support these in the kernel (even as of 2.6.12, only
    O_DIRECT files are supported) and even if, it would only allow aio_read
    and write, not open, stat and so on.

    Instead, in this module a number of (non-posix) threads are started that
    execute your read/writes and signal their completion. You don't need
    thread support in your libc or perl, and the threads created by this
    module will not be visible to the pthreads library.

    NOTICE: the threads created by this module will automatically be killed
    when the thread calling min_parallel exits. Make sure you only ever call
    min_parallel from the same thread that loaded this module.

    Although the module will work with in the presence of other threads, it
    is not reentrant, so use appropriate locking yourself.

  API NOTES
    All the "aio_*" calls are more or less thin wrappers around the syscall
    with the same name (sans "aio_"). The arguments are similar or
    identical, and they all accept an additional $callback argument which
    must be a code reference. This code reference will get called with the
    syscall return code (e.g. most syscalls return -1 on error, unlike perl,
    which usually delivers "false") as it's sole argument when the given
    syscall has been executed asynchronously.

    All functions that expect a filehandle will also accept a file
    descriptor.

    The filenames you pass to these routines *must* be absolute. The reason
    is that at the time the request is being executed, the current working
    directory could have changed. Alternatively, you can make sure that you
    never change the current working directory.

    Linux::AIO::min_parallel $nthreads
        Set the minimum number of AIO threads to $nthreads. The default is
        1, which means a single asynchronous operation can be done at one
        time (the number of outstanding operations, however, is unlimited).

        It is recommended to keep the number of threads low, as some linux
        kernel versions will scale negatively with the number of threads
        (higher parallelity => MUCH higher latency).

        Under normal circumstances you don't need to call this function, as
        this module automatically starts a single async thread.

    Linux::AIO::max_parallel $nthreads
        Sets the maximum number of AIO threads to $nthreads. If more than
        the specified number of threads are currently running, kill them.
        This function blocks until the limit is reached.

        This module automatically runs "max_parallel 0" at program end, to
        ensure that all threads are killed and that there are no outstanding
        requests.

        Under normal circumstances you don't need to call this function.

    $fileno = Linux::AIO::poll_fileno
        Return the *request result pipe filehandle*. This filehandle must be
        polled for reading by some mechanism outside this module (e.g. Event
        or select, see below). If the pipe becomes readable you have to call
        "poll_cb" to check the results.

        See "poll_cb" for an example.

    Linux::AIO::poll_cb
        Process all outstanding events on the result pipe. You have to call
        this regularly. Returns the number of events processed. Returns
        immediately when no events are outstanding.

        You can use Event to multiplex, e.g.:

           Event->io (fd => Linux::AIO::poll_fileno,
                      poll => 'r', async => 1,
                      cb => \&Linux::AIO::poll_cb);

    Linux::AIO::poll_wait
        Wait till the result filehandle becomes ready for reading (simply
        does a select on the filehandle. This is useful if you want to
        synchronously wait for some requests to finish).

        See "nreqs" for an example.

    Linux::AIO::nreqs
        Returns the number of requests currently outstanding.

        Example: wait till there are no outstanding requests anymore:

           Linux::AIO::poll_wait, Linux::AIO::poll_cb
              while Linux::AIO::nreqs;

    aio_open $pathname, $flags, $mode, $callback
        Asynchronously open or create a file and call the callback with the
        filedescriptor (NOT a perl filehandle, sorry for that, but watch
        out, this might change in the future).

        The pathname passed to "aio_open" must be absolute. See API NOTES,
        above, for an explanation.

        The $mode argument is a bitmask. See the "Fcntl" module for a list.
        They are the same as used in "sysopen".

        Example:

           aio_open "/etc/passwd", O_RDONLY, 0, sub {
              if ($_[0] >= 0) {
                 open my $fh, "<&=$_[0]";
                 print "open successful, fh is $fh\n";
                 ...
              } else {
                 die "open failed: $!\n";
              }
           };

    aio_close $fh, $callback
        Asynchronously close a file and call the callback with the result
        code.

    aio_read $fh,$offset,$length, $data,$dataoffset,$callback
    aio_write $fh,$offset,$length, $data,$dataoffset,$callback
        Reads or writes "length" bytes from the specified "fh" and "offset"
        into the scalar given by "data" and offset "dataoffset" and calls
        the callback without the actual number of bytes read (or -1 on
        error, just like the syscall).

        Example: Read 15 bytes at offset 7 into scalar $buffer, strating at
        offset 0 within the scalar:

           aio_read $fh, 7, 15, $buffer, 0, sub {
              $_[0] >= 0 or die "read error: $!";
              print "read <$buffer>\n";
           };

    aio_readahead $fh,$offset,$length, $callback
        Asynchronously reads the specified byte range into the page cache,
        using the "readahead" syscall.

        readahead() populates the page cache with data from a file so that
        subsequent reads from that file will not block on disk I/O. The
        $offset argument specifies the starting point from which data is to
        be read and $length specifies the number of bytes to be read. I/O is
        performed in whole pages, so that offset is effectively rounded down
        to a page boundary and bytes are read up to the next page boundary
        greater than or equal to (off-set+length). aio_readahead() does not
        read beyond the end of the file. The current file offset of the file
        is left unchanged.

    aio_stat $fh_or_path, $callback
    aio_lstat $fh, $callback
        Works like perl's "stat" or "lstat" in void context. The callback
        will be called after the stat and the results will be available
        using "stat _" or "-s _" etc...

        The pathname passed to "aio_stat" must be absolute. See API NOTES,
        above, for an explanation.

        Currently, the stats are always 64-bit-stats, i.e. instead of
        returning an error when stat'ing a large file, the results will be
        silently truncated unless perl itself is compiled with large file
        support.

        Example: Print the length of /etc/passwd:

           aio_stat "/etc/passwd", sub {
              $_[0] and die "stat failed: $!";
              print "size is ", -s _, "\n";
           };

    aio_unlink $pathname, $callback
        Asynchronously unlink (delete) a file and call the callback with the
        result code.

    aio_fsync $fh, $callback
        Asynchronously call fsync on the given filehandle and call the
        callback with the fsync result code.

    aio_fdatasync $fh, $callback
        Asynchronously call fdatasync on the given filehandle and call the
        callback with the fdatasync result code.

BUGS
    This module has been extensively tested in a large and very busy
    webserver for many years now.

       - aio_open gives a fd, but all other functions expect a perl filehandle.

SEE ALSO
    Coro, IO::AIO.

AUTHOR
     Marc Lehmann <schmorp@schmorp.de>
     http://home.schmorp.de/

Revision:	1.6
Committed:	Tue Jul 12 10:51:09 2005 UTC (18 years, 10 months ago) by root
Branch:	MAIN
Changes since 1.5:	+1 -2 lines
Log Message:	* empty log message *
#	User	Rev	Content
1	root	1.1	NAME
2			Linux::AIO - linux-specific aio implemented using clone
3
4			SYNOPSIS
5			use Linux::AIO;
6
7			DESCRIPTION
8	root	1.3	This module implements asynchronous I/O using the means available to
9			Linux - clone. It does not hook into the POSIX aio_* functions because
10			Linux does not yet support these in the kernel (even as of 2.6.12, only
11			O_DIRECT files are supported) and even if, it would only allow aio_read
12			and write, not open, stat and so on.
13	root	1.1
14			Instead, in this module a number of (non-posix) threads are started that
15			execute your read/writes and signal their completion. You don't need
16			thread support in your libc or perl, and the threads created by this
17			module will not be visible to the pthreads library.
18
19			NOTICE: the threads created by this module will automatically be killed
20			when the thread calling min_parallel exits. Make sure you only ever call
21			min_parallel from the same thread that loaded this module.
22
23	root	1.3	Although the module will work with in the presence of other threads, it
24			is not reentrant, so use appropriate locking yourself.
25	root	1.1
26	root	1.4	API NOTES
27			All the "aio_*" calls are more or less thin wrappers around the syscall
28			with the same name (sans "aio_"). The arguments are similar or
29			identical, and they all accept an additional $callback argument which
30			must be a code reference. This code reference will get called with the
31			syscall return code (e.g. most syscalls return -1 on error, unlike perl,
32			which usually delivers "false") as it's sole argument when the given
33			syscall has been executed asynchronously.
34
35			All functions that expect a filehandle will also accept a file
36			descriptor.
37
38	root	1.5	The filenames you pass to these routines must be absolute. The reason
39			is that at the time the request is being executed, the current working
40			directory could have changed. Alternatively, you can make sure that you
41			never change the current working directory.
42
43	root	1.1	Linux::AIO::min_parallel $nthreads
44			Set the minimum number of AIO threads to $nthreads. The default is
45			1, which means a single asynchronous operation can be done at one
46			time (the number of outstanding operations, however, is unlimited).
47
48	root	1.2	It is recommended to keep the number of threads low, as some linux
49	root	1.1	kernel versions will scale negatively with the number of threads
50			(higher parallelity => MUCH higher latency).
51
52	root	1.4	Under normal circumstances you don't need to call this function, as
53			this module automatically starts a single async thread.
54
55	root	1.2	Linux::AIO::max_parallel $nthreads
56			Sets the maximum number of AIO threads to $nthreads. If more than
57			the specified number of threads are currently running, kill them.
58			This function blocks until the limit is reached.
59
60			This module automatically runs "max_parallel 0" at program end, to
61			ensure that all threads are killed and that there are no outstanding
62			requests.
63
64	root	1.4	Under normal circumstances you don't need to call this function.
65
66	root	1.1	$fileno = Linux::AIO::poll_fileno
67			Return the request result pipe filehandle. This filehandle must be
68			polled for reading by some mechanism outside this module (e.g. Event
69			or select, see below). If the pipe becomes readable you have to call
70			"poll_cb" to check the results.
71
72	root	1.4	See "poll_cb" for an example.
73
74	root	1.1	Linux::AIO::poll_cb
75			Process all outstanding events on the result pipe. You have to call
76			this regularly. Returns the number of events processed. Returns
77			immediately when no events are outstanding.
78
79			You can use Event to multiplex, e.g.:
80
81			Event->io (fd => Linux::AIO::poll_fileno,
82			poll => 'r', async => 1,
83	root	1.4	cb => \&Linux::AIO::poll_cb);
84	root	1.1
85	root	1.2	Linux::AIO::poll_wait
86			Wait till the result filehandle becomes ready for reading (simply
87			does a select on the filehandle. This is useful if you want to
88			synchronously wait for some requests to finish).
89
90	root	1.4	See "nreqs" for an example.
91
92	root	1.1	Linux::AIO::nreqs
93			Returns the number of requests currently outstanding.
94
95	root	1.4	Example: wait till there are no outstanding requests anymore:
96
97	root	1.5	Linux::AIO::poll_wait, Linux::AIO::poll_cb
98			while Linux::AIO::nreqs;
99	root	1.4
100	root	1.1	aio_open $pathname, $flags, $mode, $callback
101			Asynchronously open or create a file and call the callback with the
102			filedescriptor (NOT a perl filehandle, sorry for that, but watch
103			out, this might change in the future).
104
105	root	1.5	The pathname passed to "aio_open" must be absolute. See API NOTES,
106			above, for an explanation.
107
108	root	1.4	The $mode argument is a bitmask. See the "Fcntl" module for a list.
109			They are the same as used in "sysopen".
110
111			Example:
112
113			aio_open "/etc/passwd", O_RDONLY, 0, sub {
114			if ($_[0] >= 0) {
115	root	1.6	open my $fh, "<&=$_[0]";
116	root	1.4	print "open successful, fh is $fh\n";
117			...
118			} else {
119			die "open failed: $!\n";
120			}
121			};
122
123	root	1.1	aio_close $fh, $callback
124			Asynchronously close a file and call the callback with the result
125			code.
126
127			aio_read $fh,$offset,$length, $data,$dataoffset,$callback
128			aio_write $fh,$offset,$length, $data,$dataoffset,$callback
129			Reads or writes "length" bytes from the specified "fh" and "offset"
130			into the scalar given by "data" and offset "dataoffset" and calls
131	root	1.4	the callback without the actual number of bytes read (or -1 on
132			error, just like the syscall).
133
134			Example: Read 15 bytes at offset 7 into scalar $buffer, strating at
135			offset 0 within the scalar:
136
137			aio_read $fh, 7, 15, $buffer, 0, sub {
138			$_[0] >= 0 or die "read error: $!";
139			print "read <$buffer>\n";
140			};
141
142			aio_readahead $fh,$offset,$length, $callback
143			Asynchronously reads the specified byte range into the page cache,
144			using the "readahead" syscall.
145
146			readahead() populates the page cache with data from a file so that
147			subsequent reads from that file will not block on disk I/O. The
148			$offset argument specifies the starting point from which data is to
149			be read and $length specifies the number of bytes to be read. I/O is
150			performed in whole pages, so that offset is effectively rounded down
151			to a page boundary and bytes are read up to the next page boundary
152			greater than or equal to (off-set+length). aio_readahead() does not
153			read beyond the end of the file. The current file offset of the file
154			is left unchanged.
155	root	1.1
156			aio_stat $fh_or_path, $callback
157			aio_lstat $fh, $callback
158			Works like perl's "stat" or "lstat" in void context. The callback
159			will be called after the stat and the results will be available
160			using "stat _" or "-s _" etc...
161
162	root	1.5	The pathname passed to "aio_stat" must be absolute. See API NOTES,
163			above, for an explanation.
164
165	root	1.1	Currently, the stats are always 64-bit-stats, i.e. instead of
166			returning an error when stat'ing a large file, the results will be
167			silently truncated unless perl itself is compiled with large file
168			support.
169
170	root	1.4	Example: Print the length of /etc/passwd:
171
172			aio_stat "/etc/passwd", sub {
173			$_[0] and die "stat failed: $!";
174			print "size is ", -s _, "\n";
175			};
176
177	root	1.1	aio_unlink $pathname, $callback
178	root	1.4	Asynchronously unlink (delete) a file and call the callback with the
179			result code.
180
181			aio_fsync $fh, $callback
182			Asynchronously call fsync on the given filehandle and call the
183			callback with the fsync result code.
184
185			aio_fdatasync $fh, $callback
186			Asynchronously call fdatasync on the given filehandle and call the
187			callback with the fdatasync result code.
188	root	1.1
189			BUGS
190			This module has been extensively tested in a large and very busy
191			webserver for many years now.
192
193			- aio_open gives a fd, but all other functions expect a perl filehandle.
194
195			SEE ALSO
196	root	1.5	Coro, IO::AIO.
197	root	1.1
198			AUTHOR
199			Marc Lehmann <schmorp@schmorp.de>
200			http://home.schmorp.de/
201