man-pages/man2/userfaultfd.2

.\" Copyright (c) 2016, IBM Corporation.
.\" Written by Mike Rapoport <rppt@linux.vnet.ibm.com>
.\"
.\" %%%LICENSE_START(VERBATIM)
.\" Permission is granted to make and distribute verbatim copies of this
.\" manual provided the copyright notice and this permission notice are
.\" preserved on all copies.
.\"
.\" Permission is granted to copy and distribute modified versions of this
.\" manual under the conditions for verbatim copying, provided that the
.\" entire resulting derived work is distributed under the terms of a
.\" permission notice identical to this one.
.\"
.\" Since the Linux kernel and libraries are constantly changing, this
.\" manual page may be incorrect or out-of-date.  The author(s) assume no
.\" responsibility for errors or omissions, or for damages resulting from
.\" the use of the information contained herein.  The author(s) may not
.\" have taken the same level of care in the production of this manual,
.\" which is licensed free of charge, as they might when working
.\" professionally.
.\"
.\" Formatted or processed versions of this manual, if unaccompanied by
.\" the source, must acknowledge the copyright and authors of this work.
.\" %%%LICENSE_END
.\"
.\" FIXME Need to mention poll/select/epoll
.\"
.TH USERFAULTFD 2 2016-12-12 "Linux" "Linux Programmer's Manual"
.SH NAME
userfaultfd \- create a file descriptor for handling page faults in user
space
.SH SYNOPSIS
.nf
.B #include <sys/types.h>
.sp
.BI "int userfaultfd(int " flags );
.fi
.PP
.IR Note :
There is no glibc wrapper for this system call; see NOTES.
.SH DESCRIPTION
.BR userfaultfd ()
creates a new userfaultfd object that can be used for delegation of page-fault
handling to a user-space application,
and returns a file descriptor that refers to the new object.
The new userfaultfd object is configured using
.BR ioctl (2).

Once the userfaultfd object is configured, the application can use
.BR read (2)
to receive userfaultfd notifications.
The reads from userfaultfd may be blocking or non-blocking,
depending on the value of
.I flags
used for the creation of the userfaultfd or subsequent calls to
.BR fcntl (2).

The following values may be bitwise ORed in
.IR flags
to change the behavior of
.BR userfaultfd ():
.TP
.BR O_CLOEXEC
Enable the close-on-exec flag for the new userfaultfd file descriptor.
See the description of the
.B O_CLOEXEC
flag in
.BR open (2).
.TP
.BR O_NONBLOCK
Enables non-blocking operation for the userfaultfd object.
See the description of the
.BR O_NONBLOCK
flag in
.BR open (2).
.\"
.SS Userfaultfd operation
After the userfaultfd object is created with
.BR userfaultfd (),
the application must enable it using the
.B UFFDIO_API
.BR ioctl (2)
operation.
This operation allows a handshake between the kernel and user space
to determine the API version and supported features.
After a successful
.B UFFDIO_API
operation,
the application then registers memory address ranges using the
.B UFFDIO_REGISTER
.BR ioctl (2)
operation.
After successful completion of a
.B UFFDIO_REGISTER
operation,
a page fault occurring in the requested memory range, and satisfying
the mode defined at the registration time, will be forwarded by the kernel to
the user-space application.
The application can then use the
.B UFFDIO_COPY
or
.B UFFDIO_ZERO
.BR ioctl (2)
operations to resolve the page fault.
.PP
Currently, userfaultfd can be used only with anonymous private memory
mappings.
.\"
.SS Configuration ioctl(2) operations
The
.BR ioctl (2)
operations described below are used to configure userfaultfd behavior.
They allow the caller to choose what features will be enabled and
what kinds of events will be delivered to the application.
.TP
.BR "UFFDIO_API	struct uffdio_api *" argp
Enable operation of the userfaultfd and perform API handshake.
The
.I uffdio_api
structure is defined as:
.in +4n
.nf

struct uffdio_api {
    __u64 api;
    __u64 features;
    __u64 ioctls;
};

.fi
.in
The
.I api
field denotes the API version requested by the application.
The kernel verifies that it can support the requested version, and sets the
.I features
and
.I ioctls
fields to bit masks representing all the available features and the generic
.BR ioctl (2
operationss available.
.\" FIXME We need to say more about the list of bits that can appear in
.\" these two fields.
.\"
.TP
.BI "UFFDIO_REGISTER	struct uffdio_register *" argp
Register a memory address range with the userfaultfd object.
The
.I uffdio_register
structure is defined as:
.in +4n
.nf

struct uffdio_range {
    __u64 start;
    __u64 end;
};

struct uffdio_register {
    struct uffdio_range range;
    __u64 mode;
    __u64 ioctls;
};

.fi
.in

The
.I range
field defines a memory range starting at
.I start
and ending at
.I end
that should be handled by the userfaultfd.

The
.I mode
field defines the mode of operation desired for this memory region.
The following values may be bitwise ORed to set the userfaultfd mode for
the specified range:

.RS
.TP
.B UFFDIO_REGISTER_MODE_MISSING
Track page faults on missing pages
.TP
.B UFFDIO_REGISTER_MODE_WP
Track page faults on write-protected pages.
Currently, the only supported mode is
.BR UFFDIO_REGISTER_MODE_MISSING .
.RE
.IP
.\" FIXME In the following, what does "answers" mean, and what are the bits?
.\" (we need a list of the bits here).
The kernel answers which ioctl commands are available for the requested
range in the
.I ioctls
field.
.\"
.TP
.BI "UFFDIO_UNREGISTER	struct uffdio_register *" argp
Unregister a memory range from userfaultfd.
.\"
.SS Range ioctl(2) operations
The range
.BR ioctl (2)
operations enable the calling application to resolve page fault
events in a consistent way.
.\" FIXME What does "consistent" mean?
.TP
.BI "UFFDIO_COPY struct uffdio_copy *" argp
Atomically copy a continuous memory chunk into the userfault registered
range and optionally wake up the blocked thread.
The source and destination addresses and the number of bytes to copy are
specified by the
.IR src ", " dst ", and " len
fields of
.IR "struct uffdio_copy" :

.in +4n
.nf
struct uffdio_copy {
    __u64 dst;
    __u64 src;
    __u64 len;
    __u64 mode;
    __s64 copy;
};
.nf
.fi
.IP
The following values may be bitwise ORed in
.IR mode
to change the behavior of the
.B UFFDIO_COPY
operation:

.RS
.TP
.B UFFDIO_COPY_MODE_DONTWAKE
Do not wake up the thread that waits for page fault resolution
.RE
.IP
The
.I copy
field of the
.I uffdio_copy
structure is used by the kernel to return the number of bytes
that was actually copied, or an error.
If
.I uffdio_copy.copy
doesn't match the
.I uffdio_copy.len
passed in input to
.BR UFFDIO_COPY ,
the operation will return
.\" FIXME In the 'copy' field? (This isn't clear.)
.BR \-EAGAIN .
If
.BR ioctl (2)
returns zero it means it succeeded, no error was reported and
the entire area was copied.
If an invalid fault happens while writing to the
.I uffdio_copy.copy
field, the system call will return
.\" FIXME In the 'copy' field? (This isn't clear.)
.BR \-EFAULT .
.I uffdio_copy.copy
is an output-only field;
it is not read by the
.B UFFDIO_COPY
operation.
.\"
.TP
.BI "UFFDIO_ZERO struct uffdio_zero *" argp
Zero out a part of memory range registered with userfaultfd.
The requested range is specified by the
.I range
field of the
.I uffdio_zeropage
structure:

.in +4n
.nf
struct uffdio_zeropage {
    struct uffdio_range range;
    __u64 mode;
    __s64 zeropage;
};
.nf
.fi
.IP
The following values may be bitwise ORed in
.IR mode
to change the behavior of
.B UFFDIO_ZERO
operation:

.RS
.TP
.B UFFDIO_ZEROPAGE_MODE_DONTWAKE
Do not wake up the thread that waits for page-fault resolution.
.RE
.IP
The
.I zeropage
field of the
.I uffdio_zero
structure is used by the kernel to return the number of bytes
that was actually zeroed,
or an error in the same manner as
.IR uffdio_copy.copy .
.\"
.TP
.BI "UFFDIO_WAKE struct uffdio_range *" argp
Wake up the thread waiting for page-fault resolution.
.SH RETURN VALUE
On success,
.BR userfaultfd ()
returns a new file descriptor that refers to the userfaultfd object.
On error, \-1 is returned, and
.I errno
is set appropriately.
.SH ERRORS
.TP
.B EINVAL
An unsupported value was specified in
.IR flags .
.TP
.BR EMFILE
The per-process limit on the number of open file descriptors has been
reached
.TP
.B ENFILE
The system-wide limit on the total number of open files has been
reached.
.TP
.B ENOMEM
Insufficient kernel memory was available.
.SH CONFORMING TO
.BR userfaultfd ()
is Linux-specific and should not be used in programs intended to be
portable.
.SH NOTES
Glibc does not provide a wrapper for this system call; call it using
.BR syscall (2).
.SH SEE ALSO
.BR fcntl (2),
.BR ioctl (2),
.BR mmap (2)

.IR Documentation/vm/userfaultfd.txt
in the Linux kernel source tree