man2/modify_ldt.2

   1 .\" Copyright (c) 1995 Michael Chastain (mec@duracef.shout.net), 22 July 1995.
   2 .\" Copyright (c) 2015 Andrew Lutomirski
   3 .\"
   4 .\" %%%LICENSE_START(GPLv2+_DOC_FULL)
   5 .\" This is free documentation; you can redistribute it and/or
   6 .\" modify it under the terms of the GNU General Public License as
   7 .\" published by the Free Software Foundation; either version 2 of
   8 .\" the License, or (at your option) any later version.
   9 .\"
  10 .\" The GNU General Public License's references to "object code"
  11 .\" and "executables" are to be interpreted as the output of any
  12 .\" document formatting or typesetting system, including
  13 .\" intermediate and printed output.
  14 .\"
  15 .\" This manual is distributed in the hope that it will be useful,
  16 .\" but WITHOUT ANY WARRANTY; without even the implied warranty of
  17 .\" MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
  18 .\" GNU General Public License for more details.
  19 .\"
  20 .\" You should have received a copy of the GNU General Public
  21 .\" License along with this manual; if not, see
  22 .\" <http://www.gnu.org/licenses/>.
  23 .\" %%%LICENSE_END
  24 .\"
  25 .TH MODIFY_LDT 2 2017-09-15 "Linux" "Linux Programmer's Manual"
  26 .SH NAME
  27 modify_ldt \- get or set a per-process LDT entry
  28 .SH SYNOPSIS
  29 .nf
  30 .B #include <sys/types.h>
  31 .PP
  32 .BI "int modify_ldt(int " func ", void *" ptr ", unsigned long " bytecount );
  33 .fi
  34 .PP
  35 .IR Note :
  36 There is no glibc wrapper for this system call; see NOTES.
  37 .SH DESCRIPTION
  38 .BR modify_ldt ()
  39 reads or writes the local descriptor table (LDT) for a process.
  40 The LDT
  41 is an array of segment descriptors that can be referenced by user code.
  42 Linux allows processes to configure a per-process (actually per-mm) LDT.
  43 For more information about the LDT, see the Intel Software Developer's
  44 Manual or the AMD Architecture Programming Manual.
  45 .PP
  46 When
  47 .I func
  48 is 0,
  49 .BR modify_ldt ()
  50 reads the LDT into the memory pointed to by
  51 .IR ptr .
  52 The number of bytes read is the smaller of
  53 .I bytecount
  54 and the actual size of the LDT, although the kernel may act as though
  55 the LDT is padded with additional trailing zero bytes.
  56 On success,
  57 .BR modify_ldt ()
  58 will return the number of bytes read.
  59 .PP
  60 When
  61 .I func
  62 is 1 or 0x11,
  63 .BR modify_ldt ()
  64 modifies the LDT entry indicated by
  65 .IR ptr\->entry_number .
  66 .I ptr
  67 points to a
  68 .I user_desc
  69 structure
  70 and
  71 .I bytecount
  72 must equal the size of this structure.
  73 .PP
  74 The
  75 .I user_desc
  76 structure is defined in \fI<asm/ldt.h>\fP as:
  77 .PP
  78 .in +4n
  79 .EX
  80 struct user_desc {
  81     unsigned int  entry_number;
  82     unsigned long base_addr;
  83     unsigned int  limit;
  84     unsigned int  seg_32bit:1;
  85     unsigned int  contents:2;
  86     unsigned int  read_exec_only:1;
  87     unsigned int  limit_in_pages:1;
  88     unsigned int  seg_not_present:1;
  89     unsigned int  useable:1;
  90 };
  91 .EE
  92 .in
  93 .PP
  94 In Linux 2.4 and earlier, this structure was named
  95 .IR modify_ldt_ldt_s .
  96 .PP
  97 The
  98 .I contents
  99 field is the segment type (data, expand-down data, non-conforming code, or
 100 conforming code).
 101 The other fields match their descriptions in the CPU manual, although
 102 .BR modify_ldt ()
 103 cannot set the hardware-defined "accessed" bit described in the CPU manual.
 104 .PP
 105 A
 106 .I user_desc
 107 is considered "empty" if
 108 .I read_exec_only
 109 and
 110 .I seg_not_present
 111 are set to 1 and all of the other fields are 0.
 112 An LDT entry can be cleared by setting it to an "empty"
 113 .I user_desc
 114 or, if
 115 .I func
 116 is 1, by setting both
 117 .I base
 118 and
 119 .I limit
 120 to 0.
 121 .PP
 122 A conforming code segment (i.e., one with
 123 .IR contents==3 )
 124 will be rejected if
 125 .I
 126 func
 127 is 1 or if
 128 .I seg_not_present
 129 is 0.
 130 .PP
 131 When
 132 .I func
 133 is 2,
 134 .BR modify_ldt ()
 135 will read zeros.
 136 This appears to be a leftover from Linux 2.4.
 137 .SH RETURN VALUE
 138 On success,
 139 .BR modify_ldt ()
 140 returns either the actual number of bytes read (for reading)
 141 or 0 (for writing).
 142 On failure,
 143 .BR modify_ldt ()
 144 returns \-1 and sets
 145 .I errno
 146 to indicate the error.
 147 .SH ERRORS
 148 .TP
 149 .B EFAULT
 150 .I ptr
 151 points outside the address space.
 152 .TP
 153 .B EINVAL
 154 .I ptr
 155 is 0,
 156 or
 157 .I func
 158 is 1 and
 159 .I bytecount
 160 is not equal to the size of the structure
 161 .IR user_desc ,
 162 or
 163 .I func
 164 is 1 or 0x11 and the new LDT entry has invalid values.
 165 .TP
 166 .B ENOSYS
 167 .I func
 168 is neither 0, 1, 2, nor 0x11.
 169 .SH CONFORMING TO
 170 This call is Linux-specific and should not be used in programs intended
 171 to be portable.
 172 .SH NOTES
 173 Glibc does not provide a wrapper for this system call; call it using
 174 .BR syscall (2).
 175 .PP
 176 .BR modify_ldt ()
 177 should not be used for thread-local storage, as it slows down context
 178 switches and only supports a limited number of threads.
 179 Threading libraries should use
 180 .BR set_thread_area (2)
 181 or
 182 .BR arch_prctl (2)
 183 instead, except on extremely old kernels that do not support those system
 184 calls.
 185 .PP
 186 The normal use for
 187 .BR modify_ldt ()
 188 is to run legacy 16-bit or segmented 32-bit code.
 189 Not all kernels allow 16-bit segments to be installed, however.
 190 .PP
 191 Even on 64-bit kernels,
 192 .BR modify_ldt ()
 193 cannot be used to create a long mode (i.e., 64-bit) code segment.
 194 The undocumented field "lm" in
 195 .IR user_desc
 196 is not useful, and, despite its name,
 197 does not result in a long mode segment.
 198 .SH BUGS
 199 On 64-bit kernels before Linux 3.19,
 200 .\" commit e30ab185c490e9a9381385529e0fd32f0a399495
 201 setting the "lm" bit in
 202 .IR user_desc
 203 prevents the descriptor from being considered empty.
 204 Keep in mind that the
 205 "lm" bit does not exist in the 32-bit headers, but these buggy kernels
 206 will still notice the bit even when set in a 32-bit process.
 207 .SH SEE ALSO
 208 .BR arch_prctl (2),
 209 .BR set_thread_area (2),
 210 .BR vm86 (2)