Computer Architecture and Operating Systems

Lecture 4: Linking and Loading

Andrei Tatarnikov
atatarnikov@hse.ru
@andrewt0301
 Example C Program
int sum(int *a, int n); int sum(int *a, int n)
{
int array[2] = {1, 2}; int i, s = 0;

int main() for (i = 0; i < n; i++) {

{ s += a[i];
int val = sum(array, 2); }
return val; return s;
} }

main.c sum.c
2
 Static Linking
 Programs are translated and linked using a compiler driver:
 linux> gcc -Og -o prog main.c sum.c
 linux> ./prog
main.c sum.c
Source files
Translators Translators
(cpp, cc1, as) (cpp, cc1, as)

main.o sum.o Separately compiled

relocatable object files
Linker (ld)
Fully linked executable object file
prog (contains code and data for all functions
3
defined in main.c and sum.c)
 Why Linkers?
Reason 1: Modularity
 Program can be written as a collection of smaller source
files, rather than one monolithic mass.
 Can build libraries of common functions (more on this
later)
 e.g., Math library, standard C library

4
 Why Linkers? (cont)
Reason 2: Efficiency
 Time: Separate compilation
 Change one source file, compile, and then relink.
 No need to recompile other source files.
Space: Libraries
 Common functions can be aggregated into a single
file...
 Yet executable files and running memory images
contain only code for the functions they actually use.

5
 What Do Linkers Do?
Step 1: Symbol resolution
 Programs define and reference symbols (global variables
and functions):
 void swap() {…} /* define symbol swap */
 swap(); /* reference symbol swap */
 int *xp = &x; /* define symbol xp, reference x */
 Symbol definitions are stored in object file (by assembler)
in symbol table.
 Symbol table is an array of structs
 Each entry includes name, size, and location of symbol.
 During symbol resolution step, the linker associates each
symbol reference with exactly one symbol definition. 6
 What Do Linkers Do? (cont)
Step 2: Relocation
 Merges separate code and data sections into single
sections
 Relocates symbols from their relative locations in the .o
files to their final absolute memory locations in the
executable.
 Updates all references to these symbols to reflect their
new positions.
Let’s look at these two steps in more detail….
7
 Three Kinds of Object Files (Modules)
Relocatable object file (.o file)
 Contains code and data in a form that can be combined
with other relocatable object files to form executable
object file.
 Each .o file is produced from exactly one source (.c) file
Executable object file (a.out file)
 Contains code and data in a form that can be copied
directly into memory and then executed.
Shared object file (.so file)
 Special type of relocatable object file that can be loaded
into memory and linked dynamically, at either load time
or run-time.
 Called Dynamic Link Libraries (DLLs) by Windows
8
 Executable and Linkable Format (ELF)
Standard binary format for object files
One unified format for
 Relocatable object files (.o),
 Executable object files (a.out)
 Shared object files (.so)
Generic name: ELF binaries

9
 ELF Object File Format
 Elf header 0
ELF header
 Word size, byte ordering, file type (.o, exec, .so), machine
type, etc. Segment header table
(required for executables)
 Segment header table
.text section
 Page size, virtual addresses memory segments (sections),
segment sizes. .rodata section

 .text section .data section

 Code .bss section
 .rodata section .symtab section
 Read only data: jump tables, ... .rel.txt section
 .data section .rel.data section
 Initialized global variables
.debug section
 .bss section
 Uninitialized global variables Section header table
 “Block Started by Symbol”
 “Better Save Space”
 Has section header but occupies no space 10
 ELF Object File Format (cont.)
 .symtab section 0
 Symbol table ELF header
 Procedure and static variable names Segment header table
 Section names and locations (required for executables)

 .rel.text section .text section

 Relocation info for .text section .rodata section
 Addresses of instructions that will need to be .data section
modified in the executable
.bss section
 Instructions for modifying.
.symtab section
 .rel.data section
 Relocation info for .data section .rel.txt section
 Addresses of pointer data that will need to be .rel.data section
modified in the merged executable .debug section
 .debug section
 Info for symbolic debugging (gcc -g) Section header table

 Section header table

 Offsets and sizes of each section 11
 Linker Symbols
Global symbols
 Symbols defined by module m that can be referenced by other
modules.
 E.g.: non-static C functions and non-static global
variables.
External symbols
 Global symbols that are referenced by module m but defined
by some other module.
Local symbols
 Symbols that are defined and referenced exclusively by
module m.
 E.g.: C functions and global variables defined with the
static attribute.
 Local linker symbols are not local program variables
12
 Step 1: Symbol Resolution
Referencing
a global…
…that’s defined here

int sum(int *a, int n); main.c int sum(int *a, int n) sum.c
{
int array[2] = {1, 2}; int i, s = 0;

int main() for (i = 0; i < n; i++) {

{ s += a[i];
int val = sum(array, 2); }
return val; return s;
} }

Defining
a global Referencing Linker knows
Linker knows a global… nothing of i or s
nothing of val …that’s defined here
13
 Local Symbols
Local non-static C variables vs. local static C variables
 local non-static C variables: stored on the stack
 local static C variables: stored in either .bss, or .data
int f()
{
static int x = 0;
return x;
Compiler allocates space in .data for each
} definition of x

int g() Creates local symbols in the symbol table with

{ unique names, e.g., x.1 and x.2.
static int x = 1;
return x;
} 14
 How Linker Resolves Duplicate Symbol Definitions

Program symbols are either strong or weak

 Strong: procedures and initialized globals
 Weak: uninitialized globals
p1.c p2.c
strong int foo=5; int foo; weak

strong p1() { p2() { strong

} }

15
 Linker’s Symbol Rules
Rule 1: Multiple strong symbols are not allowed
 Each item can be defined only once
 Otherwise: Linker error
Rule 2: Given a strong symbol and multiple weak
symbols, choose the strong symbol
 References to the weak symbol resolve to the strong
symbol
Rule 3: If there are multiple weak symbols, pick an
arbitrary one
 Can override this with gcc –fno-common
16
 Linker Puzzles
int x;
p1() {} p1() {} Link time error: two strong symbols (p1)

int x; int x; References to x will refer to the same

p1() {} p2() {}
uninitialized int. Is this what you really want?
int x; double x;
int y; p2() {}
Writes to x in p2 might overwrite y!
p1() {} Evil!
int x=7; double x; Writes to x in p2 will overwrite y!
int y=5; p2() {}
p1() {} Nasty!

int x=7; int x; References to x will refer to the same initialized

p1() {} p2() {}
variable.
Nightmare scenario: two identical weak structs, compiled by different
17
compilers with different alignment rules.
 Global Variables
Avoid if you can

Otherwise
 Use static if you can
 Initialize if you define a global variable
 Use extern if you reference an external global variable

18
 Step 2: Relocation
Relocatable Object Files Executable Object File

System code .text 0

Headers
.data
System data System code

main()
.text
main.o
swap()
main() .text

int array[2]={1,2} .data More system code

System data
sum.o .data
int array[2]={1,2}
sum() .text
.symtab
.debug

19
 Relocation Entries
int array[2] = {1, 2}; main.c
int main()
{
int val = sum(array, 2);
return val;
}

0000000000000000 <main>: main.o

0: 48 83 ec 08 sub $0x8,%rsp
4: be 02 00 00 00 mov $0x2,%esi
9: bf 00 00 00 00 mov $0x0,%edi # %edi = &array
a: R_X86_64_32 array # Relocation entry

e: e8 00 00 00 00 callq 13 <main+0x13> # sum()

f: R_X86_64_PC32 sum-0x4 # Relocation entry
13: 48 83 c4 08 add $0x8,%rsp
17: c3 retq

Source: objdump –r –d main.o

20
 Relocated .text section
00000000004004d0 <main>:
4004d0: 48 83 ec 08 sub $0x8,%rsp
4004d4: be 02 00 00 00 mov $0x2,%esi
4004d9: bf 18 10 60 00 mov $0x601018,%edi # %edi = &array
4004de: e8 05 00 00 00 callq 4004e8 <sum> # sum()
4004e3: 48 83 c4 08 add $0x8,%rsp
4004e7: c3 retq

00000000004004e8 <sum>:
4004e8: b8 00 00 00 00 mov $0x0,%eax
4004ed: ba 00 00 00 00 mov $0x0,%edx
4004f2: eb 09 jmp 4004fd <sum+0x15>
4004f4: 48 63 ca movslq %edx,%rcx
4004f7: 03 04 8f add (%rdi,%rcx,4),%eax
4004fa: 83 c2 01 add $0x1,%edx
4004fd: 39 f2 cmp %esi,%edx
4004ff: 7c f3 jl 4004f4 <sum+0xc>
400501: f3 c3 repz retq

Using PC-relative addressing for sum(): 0x4004e8 = 0x4004e3 + 0x5

Source: objdump -dx prog 21
 Loading Executable Object Files
Memory
Executable Object File Kernel virtual memory
invisible to user
0 code
ELF header
User stack
Program header table (created at runtime)
%rsp
(required for executables)
(stack
.init section pointer)
.text section Memory-mapped region for
shared libraries
.rodata section
.data section
.bss section brk
.symtab Run-time heap
(created by malloc)
.debug
Read/write data segment Loaded
.line (.data, .bss) from
the
.strtab Read-only code segment executable
Section header table (.init, .text, .rodata) file
0x400000
(required for relocatables) Unused 22
0
 Packaging Commonly Used Functions
How to package functions commonly used by
programmers?
 Math, I/O, memory management, string manipulation,
etc.
Awkward, given the linker framework so far:
 Option 1: Put all functions into a single source file
 Programmers link big object file into their programs
 Space and time inefficient
 Option 2: Put each function in a separate source file
 Programmers explicitly link appropriate binaries into their
programs
 More efficient, but burdensome on the programmer
23
 Old-fashioned Solution: Static Libraries
Static libraries (.a archive files)
 Concatenate related relocatable object files into a single
file with an index (called an archive).
 Enhance linker so that it tries to resolve unresolved
external references by looking for the symbols in one or
more archives.
 If an archive member file resolves reference, link it into
the executable.

24
 Creating Static Libraries
atoi.c printf.c random.c

Translator Translator ... Translator

atoi.o printf.o random.o

unix> ar rs libc.a \
Archiver (ar) atoi.o printf.o … random.o

libc.a C standard library

 Archiver allows incremental updates

 Recompile function that changes and replace .o file in archive.
25
 Commonly Used Libraries
libc.a (the C standard library)
 4.6 MB archive of 1496 object files.
 I/O, memory allocation, signal handling, string handling, data and time,
random numbers, integer math
libm.a (the C math library)
 2 MB archive of 444 object files.
 floating point math (sin, cos, tan, log, exp, sqrt, …)
% ar –t libc.a | sort % ar –t libm.a | sort
… …
fork.o e_acos.o
… e_acosf.o
fprintf.o e_acosh.o
fpu_control.o e_acoshf.o
fputc.o e_acoshl.o
freopen.o e_acosl.o
fscanf.o e_asin.o
fseek.o e_asinf.o
fstab.o e_asinl.o
… … 26
 Linking with Static Libraries
libvector.a

#include <stdio.h> main2.c void addvec(int *x, int *y, addvec.c

#include "vector.h" int *z, int n) {
int i;
int x[2] = {1, 2};
int y[2] = {3, 4}; for (i = 0; i < n; i++)
int z[2]; z[i] = x[i] + y[i];
}
int main()
{
void multvec(int *x, int *y, multvec.c
int *z, int n)
addvec(x, y, z, 2); {
printf("z = [%d %d]\n”, int i;
z[0], z[1]);
return 0; for (i = 0; i < n; i++)
} z[i] = x[i] * y[i];
} 27
 Linking with Static Libraries
addvec.o multvec.o

main2.c vector.h Archiver

(ar)
Translators
(cpp, cc1, as) libvector.a libc.a Static libraries
Relocatable main2.o addvec.o printf.o and any other
object files modules called by printf.o
Linker (ld)

prog2c Fully linked

executable object file
“c” for “compile-time” 28
 Using Static Libraries
 Linker’s algorithm for resolving external references:
 Scan .o files and .a files in the command line order.
 During the scan, keep a list of the current unresolved references.
 As each new .o or .a file, obj, is encountered, try to resolve each
unresolved reference in the list against the symbols defined in obj.
 If any entries in the unresolved list at end of scan, then error.
 Problem:
 Command line order matters!
 Moral: put libraries at the end of the command line.
unix> gcc -L. libtest.o -lmine
unix> gcc -L. -lmine libtest.o
libtest.o: In function `main':
libtest.o(.text+0x4): undefined reference to `libfun'
29
 Modern Solution: Shared Libraries
Static libraries have the following disadvantages:
 Duplication in the stored executables (every function
needs libc)
 Duplication in the running executables
 Minor bug fixes of system libraries require each
application to explicitly relink
Modern solution: Shared Libraries
 Object files that contain code and data that are loaded
and linked into an application dynamically, at either load-
time or run-time
 Also called: dynamic link libraries, DLLs, .so files
30
 Shared Libraries (cont.)
 Dynamic linking can occur when executable is first loaded
and run (load-time linking)
 Common case for Linux, handled automatically by the
dynamic linker (ld-linux.so)
 Standard C library (libc.so) usually dynamically linked
 Dynamic linking can also occur after program has begun
(run-time linking).
 In Linux, this is done by calls to the dlopen() interface
 Distributing software
 High-performance web servers
 Runtime library interpositioning
 Shared library routines can be shared by multiple processes
 More on this when we learn about virtual memory
31
 Dynamic Linking at Load-time
main2.c vector.h unix> gcc -shared -o libvector.so \
addvec.c multvec.c
Translators
(cpp, cc1, as) libc.so
libvector.so
Relocatable main2.o Relocation and symbol table info
object file

Linker (ld)

Partially linked prog2l

executable object file

Loader (execve) libc.so

libvector.so

Code and data

Fully linked
executable Dynamic linker (ld-linux.so)
in memory 32
 Dynamic Linking at Run-time
#include <stdio.h>
#include <stdlib.h>
#include <dlfcn.h>

int x[2] = {1, 2};

int y[2] = {3, 4};
int z[2];

int main()
{
void *handle;
void (*addvec)(int *, int *, int *, int);
char *error;

/* Dynamically load the shared library that contains addvec() */

handle = dlopen("./libvector.so", RTLD_LAZY);
if (!handle) {
fprintf(stderr, "%s\n", dlerror());
dll.c exit(1);
} 33
 Dynamic Linking at Run-time
...

/* Get a pointer to the addvec() function we just loaded */

addvec = dlsym(handle, "addvec");
if ((error = dlerror()) != NULL) {
fprintf(stderr, "%s\n", error);
exit(1);
}

/* Now we can call addvec() just like any other function */

addvec(x, y, z, 2);
printf("z = [%d %d]\n", z[0], z[1]);

/* Unload the shared library */

if (dlclose(handle) < 0) {
fprintf(stderr, "%s\n", dlerror());
exit(1);
}
dll.c return 0;
} 34
 Linking Summary
Linking is a technique that allows programs to be
constructed from multiple object files.
Linking can happen at different times in a program’s
lifetime:
 Compile time (when a program is compiled)
 Load time (when a program is loaded into memory)
 Run time (while a program is executing)
Understanding linking can help you avoid nasty errors
and make you a better programmer.
35
 Case Study: Library Interpositioning
Library interpositioning : powerful linking technique
that allows programmers to intercept calls to arbitrary
functions
Interpositioning can occur at:
 Compile time: When the source code is compiled
 Link time: When the relocatable object files are statically
linked to form an executable object file
 Load/run time: When an executable object file is loaded
into memory, dynamically linked, and then executed.

36
 Some Interpositioning Applications
Security
 Confinement (sandboxing)
 Behind the scenes encryption
Debugging
 In 2014, two Facebook engineers debugged a treacherous 1-
year old bug in their iPhone app using interpositioning
 Code in the SPDY networking stack was writing to the wrong
location
 Solved by intercepting calls to Posix write functions (write,
writev, pwrite)
Source: Facebook engineering blog post at
https://code.facebook.com/posts/313033472212144/debugging-file-corruption-on-ios/
37
 Some Interpositioning Applications
Monitoring and Profiling
Count number of calls to functions
Characterize call sites and arguments to functions
Malloc tracing
 Detecting memory leaks
 Generating address traces

38
 Example program
#include <stdio.h>  Goal: trace the addresses and sizes
#include <malloc.h> of the allocated and freed blocks,
without breaking the program, and
int main() without modifying the source
{ code.
int *p = malloc(32);
free(p);  Three solutions: interpose on the
return(0); lib malloc and free functions
} at compile time, link time, and
load/run time.
int.c
Compile-time Interpositioning
#ifdef COMPILETIME
#include <stdio.h>
mymalloc.c
#include <malloc.h>

/* malloc wrapper function */

void *mymalloc(size_t size)
{
void *ptr = malloc(size);
printf("malloc(%d)=%p\n",
(int)size, ptr);
return ptr;
}

/* free wrapper function */

void myfree(void *ptr)
{
free(ptr);
printf("free(%p)\n", ptr);
}
#endif 40
Compile-time Interpositioning
#define malloc(size) mymalloc(size) malloc.h
#define free(ptr) myfree(ptr)

void *mymalloc(size_t size);

void myfree(void *ptr);

linux> make intc

gcc -Wall -DCOMPILETIME -c mymalloc.c
gcc -Wall -I. -o intc int.c mymalloc.o
linux> make runc
./intc
malloc(32)=0x1edc010
free(0x1edc010)
linux>

41
Link-time Interpositioning
#ifdef LINKTIME mymalloc.c
#include <stdio.h>
void *__real_malloc(size_t size);
void __real_free(void *ptr);

/* malloc wrapper function */

void *__wrap_malloc(size_t size)
{
void *ptr = __real_malloc(size); /* Call libc malloc */
printf("malloc(%d) = %p\n", (int)size, ptr);
return ptr;
}

/* free wrapper function */

void __wrap_free(void *ptr)
{
__real_free(ptr); /* Call libc free */
printf("free(%p)\n", ptr);
}
#endif 42
 Link-time Interpositioning
linux> make intl
gcc -Wall -DLINKTIME -c mymalloc.c
gcc -Wall -c int.c
gcc -Wall -Wl,--wrap,malloc -Wl,--wrap,free -o intl int.o mymalloc.o
linux> make runl
./intl
malloc(32) = 0x1aa0010
free(0x1aa0010)
linux>

The “-Wl” flag passes argument to linker, replacing each

comma with a space.
The “--wrap,malloc ” arg instructs linker to resolve
references in a special way:
 Refs to malloc should be resolved as __wrap_malloc
 Refs to __real_malloc should be resolved as malloc 43
Load/Run-time Interpositioning
#ifdef RUNTIME mymalloc.c
#define _GNU_SOURCE
#include <stdio.h>
#include <stdlib.h>
#include <dlfcn.h>

/* malloc wrapper function */

void *malloc(size_t size)
{
void *(*mallocp)(size_t size);
char *error;

mallocp = dlsym(RTLD_NEXT, "malloc"); /* Get addr of libc malloc */

if ((error = dlerror()) != NULL) {
fputs(error, stderr);
exit(1);
}
char *ptr = mallocp(size); /* Call libc malloc */
printf("malloc(%d) = %p\n", (int)size, ptr);
return ptr;
}
44
Load/Run-time Interpositioning
/* free wrapper function */ mymalloc.c
void free(void *ptr)
{
void (*freep)(void *) = NULL;
char *error;

if (!ptr)
return;

freep = dlsym(RTLD_NEXT, "free"); /* Get address of libc free */

if ((error = dlerror()) != NULL) {
fputs(error, stderr);
exit(1);
}
freep(ptr); /* Call libc free */
printf("free(%p)\n", ptr);
}
#endif

45
 Load/Run-time Interpositioning
linux> make intr
gcc -Wall -DRUNTIME -shared -fpic -o mymalloc.so mymalloc.c -ldl
gcc -Wall -o intr int.c
linux> make runr
(LD_PRELOAD="./mymalloc.so" ./intr)
malloc(32) = 0xe60010
free(0xe60010)
linux>

 The LD_PRELOAD environment variable tells the

dynamic linker to resolve unresolved refs (e.g., to
malloc)by looking in mymalloc.so first.

46
 Interpositioning Recap
Compile Time
 Apparent calls to malloc/free get macro-expanded into
calls to mymalloc/myfree
Link Time
 Use linker trick to have special name resolutions
 malloc  __wrap_malloc
 __real_malloc  malloc
Load/Run Time
 Implement custom version of malloc/free that use
dynamic linking to load library malloc/free under different
names
47
Any Questions?

48