path: root/man2/userfaultfd.2

.\" Copyright (c) 2016, IBM Corporation.
.\" Written by Mike Rapoport <rppt@linux.vnet.ibm.com>
.\" and Copyright (C) 2017 Michael Kerrisk <mtk.manpages@gmail.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.\" manual page may be incorrect or out-of-date.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
.\"
.TH USERFAULTFD 2 2020-11-01 "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>
.B #include <linux/userfaultfd.h>
.PP
.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).
.PP
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).
.PP
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).
.PP
When the last file descriptor referring to a userfaultfd object is closed,
all memory ranges that were registered with the object are unregistered
and unread events are flushed.
.\"
.SS Usage
The userfaultfd mechanism is designed to allow a thread in a multithreaded
program to perform user-space paging for the other threads in the process.
When a page fault occurs for one of the regions registered
to the userfaultfd object,
the faulting thread is put to sleep and
an event is generated that can be read via the userfaultfd file descriptor.
The fault-handling thread reads events from this file descriptor and services
them using the operations described in
.BR ioctl_userfaultfd (2).
When servicing the page fault events,
the fault-handling thread can trigger a wake-up for the sleeping thread.
.PP
It is possible for the faulting threads and the fault-handling threads
to run in the context of different processes.
In this case, these threads may belong to different programs,
and the program that executes the faulting threads
will not necessarily cooperate with the program that handles the page faults.
In such non-cooperative mode,
the process that monitors userfaultfd and handles page faults
needs to be aware of the changes in the virtual memory layout
of the faulting process to avoid memory corruption.
.PP
Starting from Linux 4.11,
userfaultfd can also notify the fault-handling threads about changes
in the virtual memory layout of the faulting process.
In addition, if the faulting process invokes
.BR fork (2),
the userfaultfd objects associated with the parent may be duplicated
into the child process and the userfaultfd monitor will be notified
(via the
.B UFFD_EVENT_FORK
described below)
about the file descriptor associated with the userfault objects
created for the child process,
which allows the userfaultfd monitor to perform user-space paging
for the child process.
Unlike page faults which have to be synchronous and require an
explicit or implicit wakeup,
all other events are delivered asynchronously and
the non-cooperative process resumes execution as
soon as the userfaultfd manager executes
.BR read (2).
The userfaultfd manager should carefully synchronize calls to
.B UFFDIO_COPY
with the processing of events.
.PP
The current asynchronous model of the event delivery is optimal for
single threaded non-cooperative userfaultfd manager implementations.
.\" Regarding the preceding sentence, Mike Rapoport says:
.\"     The major point here is that current events delivery model could be
.\"     problematic for multi-threaded monitor. I even suspect that it would be
.\"     impossible to ensure synchronization between page faults and non-page
.\"     fault events in multi-threaded monitor.
.\" .PP
.\" FIXME elaborate about non-cooperating mode, describe its limitations
.\" for kernels before 4.11, features added in 4.11
.\" and limitations remaining in 4.11
.\" Maybe it's worth adding a dedicated sub-section...
.\"
.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.
This operation must be performed before any of the other
.BR ioctl (2)
operations described below (or those operations fail with the
.BR EINVAL
error).
.PP
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_ZEROPAGE
.BR ioctl (2)
operations to resolve the page fault.
.PP
Starting from Linux 4.14, if the application sets the
.B UFFD_FEATURE_SIGBUS
feature bit using the
.B UFFDIO_API
.BR ioctl (2),
no page-fault notification will be forwarded to user space.
Instead a
.B SIGBUS
signal is delivered to the faulting process.
With this feature,
userfaultfd can be used for robustness purposes to simply catch
any access to areas within the registered address range that do not
have pages allocated, without having to listen to userfaultfd events.
No userfaultfd monitor will be required for dealing with such memory
accesses.
For example, this feature can be useful for applications that
want to prevent the kernel from automatically allocating pages and filling
holes in sparse files when the hole is accessed through a memory mapping.
.PP
The
.B UFFD_FEATURE_SIGBUS
feature is implicitly inherited through
.BR fork (2)
if used in combination with
.BR UFFD_FEATURE_FORK .
.PP
Details of the various
.BR ioctl (2)
operations can be found in
.BR ioctl_userfaultfd (2).
.PP
Since Linux 4.11, events other than page-fault may enabled during
.B UFFDIO_API
operation.
.PP
Up to Linux 4.11,
userfaultfd can be used only with anonymous private memory mappings.
Since Linux 4.11,
userfaultfd can be also used with hugetlbfs and shared memory mappings.
.PP
.\"
.SS Reading from the userfaultfd structure
Each
.BR read (2)
from the userfaultfd file descriptor returns one or more
.I uffd_msg
structures, each of which describes a page-fault event
or an event required for the non-cooperative userfaultfd usage:
.PP
.in +4n
.EX
struct uffd_msg {
    __u8  event;            /* Type of event */
    ...
    union {
        struct {
            __u64 flags;    /* Flags describing fault */
            __u64 address;  /* Faulting address */
        } pagefault;

        struct {            /* Since Linux 4.11 */
            __u32 ufd;      /* Userfault file descriptor
                               of the child process */
        } fork;

        struct {            /* Since Linux 4.11 */
            __u64 from;     /* Old address of remapped area */
            __u64 to;       /* New address of remapped area */
            __u64 len;      /* Original mapping length */
        } remap;

        struct {            /* Since Linux 4.11 */
            __u64 start;    /* Start address of removed area */
            __u64 end;      /* End address of removed area */
        } remove;
        ...
    } arg;

    /* Padding fields omitted */
} __packed;
.EE
.in
.PP
If multiple events are available and the supplied buffer is large enough,
.BR read (2)
returns as many events as will fit in the supplied buffer.
If the buffer supplied to
.BR read (2)
is smaller than the size of the
.I uffd_msg
structure, the
.BR read (2)
fails with the error
.BR EINVAL .
.PP
The fields set in the
.I uffd_msg
structure are as follows:
.TP
.I event
The type of event.
Depending of the event type,
different fields of the
.I arg
union represent details required for the event processing.
The non-page-fault events are generated only when appropriate feature
is enabled during API handshake with
.B UFFDIO_API
.BR ioctl (2).
.IP
The following values can appear in the
.I event
field:
.RS
.TP
.BR UFFD_EVENT_PAGEFAULT " (since Linux 4.3)"
A page-fault event.
The page-fault details are available in the
.I pagefault
field.
.TP
.BR UFFD_EVENT_FORK " (since Linux 4.11)"
Generated when the faulting process invokes
.BR fork (2)
(or
.BR clone (2)
without the
.BR CLONE_VM
flag).
The event details are available in the
.I fork
field.
.\" FIXME describe duplication of userfault file descriptor during fork
.TP
.BR UFFD_EVENT_REMAP " (since Linux 4.11)"
Generated when the faulting process invokes
.BR mremap (2).
The event details are available in the
.I remap
field.
.TP
.BR UFFD_EVENT_REMOVE " (since Linux 4.11)"
Generated when the faulting process invokes
.BR madvise (2)
with
.BR MADV_DONTNEED
or
.BR MADV_REMOVE
advice.
The event details are available in the
.I remove
field.
.TP
.BR UFFD_EVENT_UNMAP " (since Linux 4.11)"
Generated when the faulting process unmaps a memory range,
either explicitly using
.BR munmap (2)
or implicitly during
.BR mmap (2)
or
.BR mremap (2).
The event details are available in the
.I remove
field.
.RE
.TP
.I pagefault.address
The address that triggered the page fault.
.TP
.I pagefault.flags
A bit mask of flags that describe the event.
For
.BR UFFD_EVENT_PAGEFAULT ,
the following flag may appear:
.RS
.TP
.B UFFD_PAGEFAULT_FLAG_WRITE
If the address is in a range that was registered with the
.B UFFDIO_REGISTER_MODE_MISSING
flag (see
.BR ioctl_userfaultfd (2))
and this flag is set, this a write fault;
otherwise it is a read fault.
.\"
.\" UFFD_PAGEFAULT_FLAG_WP is not yet supported.
.RE
.TP
.I fork.ufd
The file descriptor associated with the userfault object
created for the child created by
.BR fork (2).
.TP
.I remap.from
The original address of the memory range that was remapped using
.BR mremap (2).
.TP
.I remap.to
The new address of the memory range that was remapped using
.BR mremap (2).
.TP
.I remap.len
The original length of the memory range that was remapped using
.BR mremap (2).
.TP
.I remove.start
The start address of the memory range that was freed using
.BR madvise (2)
or unmapped
.TP
.I remove.end
The end address of the memory range that was freed using
.BR madvise (2)
or unmapped
.PP
A
.BR read (2)
on a userfaultfd file descriptor can fail with the following errors:
.TP
.B EINVAL
The userfaultfd object has not yet been enabled using the
.BR UFFDIO_API
.BR ioctl (2)
operation
.PP
If the
.B O_NONBLOCK
flag is enabled in the associated open file description,
the userfaultfd file descriptor can be monitored with
.BR poll (2),
.BR select (2),
and
.BR epoll (7).
When events are available, the file descriptor indicates as readable.
If the
.B O_NONBLOCK
flag is not enabled, then
.BR poll (2)
(always) indicates the file as having a
.BR POLLERR
condition, and
.BR select (2)
indicates the file descriptor as both readable and writable.
.\" FIXME What is the reason for this seemingly odd behavior with respect
.\" to the O_NONBLOCK flag? (see userfaultfd_poll() in fs/userfaultfd.c).
.\" Something needs to be said about this.
.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.
.TP
.BR EPERM " (since Linux 5.2)"
.\" cefdca0a86be517bc390fc4541e3674b8e7803b0
The caller is not privileged (does not have the
.B CAP_SYS_PTRACE
capability in the initial user namespace), and
.I /proc/sys/vm/unprivileged_userfaultfd
has the value 0.
.SH VERSIONS
The
.BR userfaultfd ()
system call first appeared in Linux 4.3.
.PP
The support for hugetlbfs and shared memory areas and
non-page-fault events was added in Linux 4.11
.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).
.PP
The userfaultfd mechanism can be used as an alternative to
traditional user-space paging techniques based on the use of the
.BR SIGSEGV
signal and
.BR mmap (2).
It can also be used to implement lazy restore
for checkpoint/restore mechanisms,
as well as post-copy migration to allow (nearly) uninterrupted execution
when transferring virtual machines and Linux containers
from one host to another.
.SH BUGS
If the
.B UFFD_FEATURE_EVENT_FORK
is enabled and a system call from the
.BR fork (2)
family is interrupted by a signal or failed, a stale userfaultfd descriptor
might be created.
In this case, a spurious
.B UFFD_EVENT_FORK
will be delivered to the userfaultfd monitor.
.SH EXAMPLES
The program below demonstrates the use of the userfaultfd mechanism.
The program creates two threads, one of which acts as the
page-fault handler for the process, for the pages in a demand-page zero
region created using
.BR mmap (2).
.PP
The program takes one command-line argument,
which is the number of pages that will be created in a mapping
whose page faults will be handled via userfaultfd.
After creating a userfaultfd object,
the program then creates an anonymous private mapping of the specified size
and registers the address range of that mapping using the
.B UFFDIO_REGISTER
.BR ioctl (2)
operation.
The program then creates a second thread that will perform the
task of handling page faults.
.PP
The main thread then walks through the pages of the mapping fetching
bytes from successive pages.
Because the pages have not yet been accessed,
the first access of a byte in each page will trigger a page-fault event
on the userfaultfd file descriptor.
.PP
Each of the page-fault events is handled by the second thread,
which sits in a loop processing input from the userfaultfd file descriptor.
In each loop iteration, the second thread first calls
.BR poll (2)
to check the state of the file descriptor,
and then reads an event from the file descriptor.
All such events should be
.B UFFD_EVENT_PAGEFAULT
events,
which the thread handles by copying a page of data into
the faulting region using the
.B UFFDIO_COPY
.BR ioctl (2)
operation.
.PP
The following is an example of what we see when running the program:
.PP
.in +4n
.EX
$ \fB./userfaultfd_demo 3\fP
Address returned by mmap() = 0x7fd30106c000

fault_handler_thread():
    poll() returns: nready = 1; POLLIN = 1; POLLERR = 0
    UFFD_EVENT_PAGEFAULT event: flags = 0; address = 7fd30106c00f
        (uffdio_copy.copy returned 4096)
Read address 0x7fd30106c00f in main(): A
Read address 0x7fd30106c40f in main(): A
Read address 0x7fd30106c80f in main(): A
Read address 0x7fd30106cc0f in main(): A

fault_handler_thread():
    poll() returns: nready = 1; POLLIN = 1; POLLERR = 0
    UFFD_EVENT_PAGEFAULT event: flags = 0; address = 7fd30106d00f
        (uffdio_copy.copy returned 4096)
Read address 0x7fd30106d00f in main(): B
Read address 0x7fd30106d40f in main(): B
Read address 0x7fd30106d80f in main(): B
Read address 0x7fd30106dc0f in main(): B

fault_handler_thread():
    poll() returns: nready = 1; POLLIN = 1; POLLERR = 0
    UFFD_EVENT_PAGEFAULT event: flags = 0; address = 7fd30106e00f
        (uffdio_copy.copy returned 4096)
Read address 0x7fd30106e00f in main(): C
Read address 0x7fd30106e40f in main(): C
Read address 0x7fd30106e80f in main(): C
Read address 0x7fd30106ec0f in main(): C
.EE
.in
.SS Program source
\&
.EX
/* userfaultfd_demo.c

   Licensed under the GNU General Public License version 2 or later.
*/
#define _GNU_SOURCE
#include <inttypes.h>
#include <sys/types.h>
#include <stdio.h>
#include <linux/userfaultfd.h>
#include <pthread.h>
#include <errno.h>
#include <unistd.h>
#include <stdlib.h>
#include <fcntl.h>
#include <signal.h>
#include <poll.h>
#include <string.h>
#include <sys/mman.h>
#include <sys/syscall.h>
#include <sys/ioctl.h>
#include <poll.h>

#define errExit(msg)    do { perror(msg); exit(EXIT_FAILURE); \e
                        } while (0)

static int page_size;

static void *
fault_handler_thread(void *arg)
{
    static struct uffd_msg msg;   /* Data read from userfaultfd */
    static int fault_cnt = 0;     /* Number of faults so far handled */
    long uffd;                    /* userfaultfd file descriptor */
    static char *page = NULL;
    struct uffdio_copy uffdio_copy;
    ssize_t nread;

    uffd = (long) arg;

    /* Create a page that will be copied into the faulting region */

    if (page == NULL) {
        page = mmap(NULL, page_size, PROT_READ | PROT_WRITE,
                    MAP_PRIVATE | MAP_ANONYMOUS, \-1, 0);
        if (page == MAP_FAILED)
            errExit("mmap");
    }

    /* Loop, handling incoming events on the userfaultfd
       file descriptor */

    for (;;) {

        /* See what poll() tells us about the userfaultfd */

        struct pollfd pollfd;
        int nready;
        pollfd.fd = uffd;
        pollfd.events = POLLIN;
        nready = poll(&pollfd, 1, \-1);
        if (nready == \-1)
            errExit("poll");

        printf("\enfault_handler_thread():\en");
        printf("    poll() returns: nready = %d; "
                "POLLIN = %d; POLLERR = %d\en", nready,
                (pollfd.revents & POLLIN) != 0,
                (pollfd.revents & POLLERR) != 0);

        /* Read an event from the userfaultfd */

        nread = read(uffd, &msg, sizeof(msg));
        if (nread == 0) {
            printf("EOF on userfaultfd!\en");
            exit(EXIT_FAILURE);
        }

        if (nread == \-1)
            errExit("read");

        /* We expect only one kind of event; verify that assumption */

        if (msg.event != UFFD_EVENT_PAGEFAULT) {
            fprintf(stderr, "Unexpected event on userfaultfd\en");
            exit(EXIT_FAILURE);
        }

        /* Display info about the page\-fault event */

        printf("    UFFD_EVENT_PAGEFAULT event: ");
        printf("flags = %"PRIx64"; ", msg.arg.pagefault.flags);
        printf("address = %"PRIx64"\en", msg.arg.pagefault.address);

        /* Copy the page pointed to by \(aqpage\(aq into the faulting
           region. Vary the contents that are copied in, so that it
           is more obvious that each fault is handled separately. */

        memset(page, \(aqA\(aq + fault_cnt % 20, page_size);
        fault_cnt++;

        uffdio_copy.src = (unsigned long) page;

        /* We need to handle page faults in units of pages(!).
           So, round faulting address down to page boundary */

        uffdio_copy.dst = (unsigned long) msg.arg.pagefault.address &
                                           \(ti(page_size \- 1);
        uffdio_copy.len = page_size;
        uffdio_copy.mode = 0;
        uffdio_copy.copy = 0;
        if (ioctl(uffd, UFFDIO_COPY, &uffdio_copy) == \-1)
            errExit("ioctl\-UFFDIO_COPY");

        printf("        (uffdio_copy.copy returned %"PRId64")\en",
                uffdio_copy.copy);
    }
}

int
main(int argc, char *argv[])
{
    long uffd;          /* userfaultfd file descriptor */
    char *addr;         /* Start of region handled by userfaultfd */
    uint64_t len;       /* Length of region handled by userfaultfd */
    pthread_t thr;      /* ID of thread that handles page faults */
    struct uffdio_api uffdio_api;
    struct uffdio_register uffdio_register;
    int s;

    if (argc != 2) {
        fprintf(stderr, "Usage: %s num\-pages\en", argv[0]);
        exit(EXIT_FAILURE);
    }

    page_size = sysconf(_SC_PAGE_SIZE);
    len = strtoull(argv[1], NULL, 0) * page_size;

    /* Create and enable userfaultfd object */

    uffd = syscall(__NR_userfaultfd, O_CLOEXEC | O_NONBLOCK);
    if (uffd == \-1)
        errExit("userfaultfd");

    uffdio_api.api = UFFD_API;
    uffdio_api.features = 0;
    if (ioctl(uffd, UFFDIO_API, &uffdio_api) == \-1)
        errExit("ioctl\-UFFDIO_API");

    /* Create a private anonymous mapping. The memory will be
       demand\-zero paged\-\-that is, not yet allocated. When we
       actually touch the memory, it will be allocated via
       the userfaultfd. */

    addr = mmap(NULL, len, PROT_READ | PROT_WRITE,
                MAP_PRIVATE | MAP_ANONYMOUS, \-1, 0);
    if (addr == MAP_FAILED)
        errExit("mmap");

    printf("Address returned by mmap() = %p\en", addr);

    /* Register the memory range of the mapping we just created for
       handling by the userfaultfd object. In mode, we request to track
       missing pages (i.e., pages that have not yet been faulted in). */

    uffdio_register.range.start = (unsigned long) addr;
    uffdio_register.range.len = len;
    uffdio_register.mode = UFFDIO_REGISTER_MODE_MISSING;
    if (ioctl(uffd, UFFDIO_REGISTER, &uffdio_register) == \-1)
        errExit("ioctl\-UFFDIO_REGISTER");

    /* Create a thread that will process the userfaultfd events */

    s = pthread_create(&thr, NULL, fault_handler_thread, (void *) uffd);
    if (s != 0) {
        errno = s;
        errExit("pthread_create");
    }

    /* Main thread now touches memory in the mapping, touching
       locations 1024 bytes apart. This will trigger userfaultfd
       events for all pages in the region. */

    int l;
    l = 0xf;    /* Ensure that faulting address is not on a page
                   boundary, in order to test that we correctly
                   handle that case in fault_handling_thread() */
    while (l < len) {
        char c = addr[l];
        printf("Read address %p in main(): ", addr + l);
        printf("%c\en", c);
        l += 1024;
        usleep(100000);         /* Slow things down a little */
    }

    exit(EXIT_SUCCESS);
}
.EE
.SH SEE ALSO
.BR fcntl (2),
.BR ioctl (2),
.BR ioctl_userfaultfd (2),
.BR madvise (2),
.BR mmap (2)
.PP
.IR Documentation/admin-guide/mm/userfaultfd.rst
in the Linux kernel source tree
.SH COLOPHON
This page is part of release 5.09 of the Linux
.I man-pages
project.
A description of the project,
information about reporting bugs,
and the latest version of this page,
can be found at
\%https://www.kernel.org/doc/man\-pages/.