man2/setfsuid.2

   1 .\" Copyright (C) 1995, Thomas K. Dyas <tdyas@eden.rutgers.edu>
   2 .\" and Copyright (C) 2013, 2019, Michael Kerrisk <mtk.manpages@gmail.com>
   3 .\"
   4 .\" %%%LICENSE_START(VERBATIM)
   5 .\" Permission is granted to make and distribute verbatim copies of this
   6 .\" manual provided the copyright notice and this permission notice are
   7 .\" preserved on all copies.
   8 .\"
   9 .\" Permission is granted to copy and distribute modified versions of this
  10 .\" manual under the conditions for verbatim copying, provided that the
  11 .\" entire resulting derived work is distributed under the terms of a
  12 .\" permission notice identical to this one.
  13 .\"
  14 .\" Since the Linux kernel and libraries are constantly changing, this
  15 .\" manual page may be incorrect or out-of-date.  The author(s) assume no
  16 .\" responsibility for errors or omissions, or for damages resulting from
  17 .\" the use of the information contained herein.  The author(s) may not
  18 .\" have taken the same level of care in the production of this manual,
  19 .\" which is licensed free of charge, as they might when working
  20 .\" professionally.
  21 .\"
  22 .\" Formatted or processed versions of this manual, if unaccompanied by
  23 .\" the source, must acknowledge the copyright and authors of this work.
  24 .\" %%%LICENSE_END
  25 .\"
  26 .\" Created   1995-08-06 Thomas K. Dyas <tdyas@eden.rutgers.edu>
  27 .\" Modified  2000-07-01 aeb
  28 .\" Modified  2002-07-23 aeb
  29 .\" Modified, 27 May 2004, Michael Kerrisk <mtk.manpages@gmail.com>
  30 .\"     Added notes on capability requirements
  31 .\"
  32 .TH SETFSUID 2 2019-05-09 "Linux" "Linux Programmer's Manual"
  33 .SH NAME
  34 setfsuid \- set user identity used for filesystem checks
  35 .SH SYNOPSIS
  36 .B #include <sys/fsuid.h>
  37 .PP
  38 .BI "int setfsuid(uid_t " fsuid );
  39 .SH DESCRIPTION
  40 On Linux, a process has both a filesystem user ID and an effective user ID.
  41 The (Linux-specific) filesystem user ID is used
  42 for permissions checking when accessing filesystem objects,
  43 while the effective user ID is used for various other kinds
  44 of permissions checks (see
  45 .BR credentials (7)).
  46 .PP
  47 Normally, the value of the process's filesystem user ID
  48 is the same as the value of its effective user ID.
  49 This is so, because whenever a process's effective user ID is changed,
  50 the kernel also changes the filesystem user ID to be the same as
  51 the new value of the effective user ID.
  52 A process can cause the value of its filesystem user ID to diverge
  53 from its effective user ID by using
  54 .BR setfsuid ()
  55 to change its filesystem user ID to the value given in
  56 .IR fsuid .
  57 .PP
  58 Explicit calls to
  59 .BR setfsuid ()
  60 and
  61 .BR setfsgid (2)
  62 are (were) usually used only by programs such as the Linux NFS server that
  63 need to change what user and group ID is used for file access without a
  64 corresponding change in the real and effective user and group IDs.
  65 A change in the normal user IDs for a program such as the NFS server
  66 is (was) a security hole that can expose it to unwanted signals.
  67 (However, this issue is historical; see below.)
  68 .PP
  69 .BR setfsuid ()
  70 will succeed only if the caller is the superuser or if
  71 .I fsuid
  72 matches either the caller's real user ID, effective user ID,
  73 saved set-user-ID, or current filesystem user ID.
  74 .SH RETURN VALUE
  75 On both success and failure,
  76 this call returns the previous filesystem user ID of the caller.
  77 .SH VERSIONS
  78 This system call is present in Linux since version 1.2.
  79 .\" This system call is present since Linux 1.1.44
  80 .\" and in libc since libc 4.7.6.
  81 .SH CONFORMING TO
  82 .BR setfsuid ()
  83 is Linux-specific and should not be used in programs intended
  84 to be portable.
  85 .SH NOTES
  86 At the time when this system call was introduced, one process
  87 could send a signal to another process with the same effective user ID.
  88 This meant that if a privileged process changed its effective user ID
  89 for the purpose of file permission checking,
  90 then it could become vulnerable to receiving signals
  91 sent by another (unprivileged) process with the same user ID.
  92 The filesystem user ID attribute was thus added to allow a process to
  93 change its user ID for the purposes of file permission checking without
  94 at the same time becoming vulnerable to receiving unwanted signals.
  95 Since Linux 2.0, signal permission handling is different (see
  96 .BR kill (2)),
  97 with the result that a process can change its effective user ID
  98 without being vulnerable to receiving signals from unwanted processes.
  99 Thus,
 100 .BR setfsuid ()
 101 is nowadays unneeded and should be avoided in new applications
 102 (likewise for
 103 .BR setfsgid (2)).
 104 .PP
 105 The original Linux
 106 .BR setfsuid ()
 107 system call supported only 16-bit user IDs.
 108 Subsequently, Linux 2.4 added
 109 .BR setfsuid32 ()
 110 supporting 32-bit IDs.
 111 The glibc
 112 .BR setfsuid ()
 113 wrapper function transparently deals with the variation across kernel versions.
 114 .SS C library/kernel differences
 115 In glibc 2.15 and earlier,
 116 when the wrapper for this system call determines that the argument can't be
 117 passed to the kernel without integer truncation (because the kernel
 118 is old and does not support 32-bit user IDs),
 119 it will return \-1 and set \fIerrno\fP to
 120 .B EINVAL
 121 without attempting
 122 the system call.
 123 .SH BUGS
 124 No error indications of any kind are returned to the caller,
 125 and the fact that both successful and unsuccessful calls return
 126 the same value makes it impossible to directly determine
 127 whether the call succeeded or failed.
 128 Instead, the caller must resort to looking at the return value
 129 from a further call such as
 130 .IR setfsuid(\-1)
 131 (which will always fail), in order to determine if a preceding call to
 132 .BR setfsuid ()
 133 changed the filesystem user ID.
 134 At the very
 135 least,
 136 .B EPERM
 137 should be returned when the call fails (because the caller lacks the
 138 .B CAP_SETUID
 139 capability).
 140 .SH SEE ALSO
 141 .BR kill (2),
 142 .BR setfsgid (2),
 143 .BR capabilities (7),
 144 .BR credentials (7)