Intel MAC assembly examples

On 2009-04-08 04:29:54 +1000, Reinder Verlinde <reinder RemoveThis @verlinde.invalid> said:

> In article <49d9f0a6$0$20975$afc38c87@news.optusnet.com.au>,
> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>
>> Does anyone know where I can find some examples of assembly programs in
>> 'as' for intel OSX
>
> Take any C or C++ source file, and do:
>
> gcc -S filename.c
>
> This produces a filename.s file containing the as translation of
> filename.c
>
> Reinder

Thanks for that gave it a try with a hello world program and I am
reviewing the output now.
--
.....
"When asked what he would like on his tomb stone he said more bass"

On 2009-04-08 22:41:17 +1000, Aaron Spiteri <azzmosphere(at)gmail.com> said:

> On 2009-04-08 04:29:54 +1000, Reinder Verlinde <reinder.TakeThisOut@verlinde.invalid> said:
>
>> In article <49d9f0a6$0$20975$afc38c87@news.optusnet.com.au>,
>> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>>
>>> Does anyone know where I can find some examples of assembly programs in
>>> 'as' for intel OSX
>>
>> Take any C or C++ source file, and do:
>>
>> gcc -S filename.c
>>
>> This produces a filename.s file containing the as translation of
>> filename.c
>>
>> Reinder
>
> Thanks for that gave it a try with a hello world program and I am
> reviewing the output now.

I wrote a small C program "Hello World" and tested that it compiled.
Produced the ASM output and got the following:

.cstring
LC0:
.ascii "Hello World\0"
.text
..globl _main
_main:
pushl %ebp
movl %esp, %ebp
pushl %ebx
subl $20, %esp
call ___i686.get_pc_thunk.bx
"L00000000001$pb":
leal LC0-"L00000000001$pb"(%ebx), %eax
movl %eax, (%esp)
call L_puts$stub
movl $0, %eax
addl $20, %esp
popl %ebx
popl %ebp
ret
.section
__IMPORT,__jump_table,symbol_stubs,self_modifying_code+pure_instructions,5
L_puts$stub:


.indirect_symbol _puts
hlt ; hlt ; hlt ; hlt ; hlt
.subsections_via_symbols
.section __TEXT,__textcoal_nt,coalesced,pure_instructions
..weak_definition ___i686.get_pc_thunk.bx
..private_extern ___i686.get_pc_thunk.bx
___i686.get_pc_thunk.bx:
movl (%esp), %ebx
ret

All though the program compiled to a executable, When I tried to link
it using the following command:

as -o hello.o hello.s
ld -o hello hello.o

I keep getting the message:

ld: Undefined symbols:
_puts

I imagine that this has something to do with the line:

L_puts$stub:

but can not understand why it compiles using gcc but not when ld is
call directly. Any idea's would be greatly appreciated.
--
.....
"When asked what he would like on his tomb stone he said more bass"

On 2009-04-08 22:41:17 +1000, Aaron Spiteri <azzmosphere(at)gmail.com> said:

> On 2009-04-08 04:29:54 +1000, Reinder Verlinde <reinder RemoveThis @verlinde.invalid> said:
>
>> In article <49d9f0a6$0$20975$afc38c87@news.optusnet.com.au>,
>> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>>
>>> Does anyone know where I can find some examples of assembly programs in
>>> 'as' for intel OSX
>>
>> Take any C or C++ source file, and do:
>>
>> gcc -S filename.c
>>
>> This produces a filename.s file containing the as translation of
>> filename.c
>>
>> Reinder
>
> Thanks for that gave it a try with a hello world program and I am
> reviewing the output now.

I have tried your suggestion and all worked well but when I went to
compile the ASM code into a executable kept on getting the message:

ld: Undefined symbols:
_puts


I was using the following commands:

as -o hello.o hello.s
ld -o hello hello.o

The code compiled correctly under C

gcc -o hello hello.c

Here is a copy:

.cstring
LC0:
.ascii "Hello World\0"
.text
..globl _main
_main:
pushl %ebp
movl %esp, %ebp
pushl %ebx
subl $20, %esp
call ___i686.get_pc_thunk.bx
"L00000000001$pb":
leal LC0-"L00000000001$pb"(%ebx), %eax
movl %eax, (%esp)
call L_puts$stub
movl $0, %eax
addl $20, %esp
popl %ebx
popl %ebp
ret
.section
__IMPORT,__jump_table,symbol_stubs,self_modifying_code+pure_instructions,5
L_puts$stub:


.indirect_symbol _puts
hlt ; hlt ; hlt ; hlt ; hlt
.subsections_via_symbols
.section __TEXT,__textcoal_nt,coalesced,pure_instructions
..weak_definition ___i686.get_pc_thunk.bx
..private_extern ___i686.get_pc_thunk.bx
___i686.get_pc_thunk.bx:
movl (%esp), %ebx
ret

I am assuming that it has something to do with the puts function but
can not figure out what. My assumption would be this is the very same
assembly that C is getting translated into. Should I be including
anything or am I missing some sort of critical option?

--
.....
"When asked what he would like on his tomb stone he said more bass"

In article <49dc9e60$0$12942$afc38c87@news.optusnet.com.au>,
Aaron Spiteri <azzmosphere(at)gmail.com> writes:
>
> All though the program compiled to a executable, When I tried to link
> it using the following command:
>
> as -o hello.o hello.s
> ld -o hello hello.o
>
> I keep getting the message:
>
> ld: Undefined symbols:
> _puts
>
> I imagine that this has something to do with the line:
>
> L_puts$stub:
>
> but can not understand why it compiles using gcc but not when ld is
> call directly. Any idea's would be greatly appreciated.

Probably because you aren't calling ld the way gcc would.
For example, you started with C, so you probably need the
C library for puts(), but ld won't know that.

Try: ld -o hello -lc hello.o

Alternatively, use the gcc command to link it, and it will
call ld with the correct parameters.

--
Andrew Gabriel
[email address is not usable -- followup in the newsgroup]

Aaron Spiteri <azzmosphere(at)gmail.com> wrote:

> On 2009-04-08 22:41:17 +1000, Aaron Spiteri <azzmosphere(at)gmail.com>
> said:
>
>> On 2009-04-08 04:29:54 +1000, Reinder Verlinde <reinder.DeleteThis@verlinde.invalid>
>> said:
>>
>>> In article <49d9f0a6$0$20975$afc38c87@news.optusnet.com.au>,
>>> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>>>
>>>> Does anyone know where I can find some examples of assembly programs in
>>>> 'as' for intel OSX
> I wrote a small C program "Hello World" and tested that it compiled.
> Produced the ASM output and got the following:
>
> .cstring
> LC0:
> .ascii "Hello World\0"
> .text
> .globl _main
> _main:
> pushl %ebp
> movl %esp, %ebp
> pushl %ebx
> subl $20, %esp
> call ___i686.get_pc_thunk.bx
> "L00000000001$pb":
> leal LC0-"L00000000001$pb"(%ebx), %eax
> movl %eax, (%esp)
> call L_puts$stub
> movl $0, %eax
> addl $20, %esp
> popl %ebx
> popl %ebp
> ret
> .section
> __IMPORT,__jump_table,symbol_stubs,self_modifying_code+pure_instructions,5
> L_puts$stub:
>
>
> .indirect_symbol _puts
> hlt ; hlt ; hlt ; hlt ; hlt
> .subsections_via_symbols
> .section __TEXT,__textcoal_nt,coalesced,pure_instructions
> .weak_definition ___i686.get_pc_thunk.bx
> .private_extern ___i686.get_pc_thunk.bx
> ___i686.get_pc_thunk.bx:
> movl (%esp), %ebx
> ret
>
> All though the program compiled to a executable, When I tried to link
> it using the following command:
>
> as -o hello.o hello.s
> ld -o hello hello.o
>
> I keep getting the message:
>
> ld: Undefined symbols:
> _puts
>
> I imagine that this has something to do with the line:
>
> L_puts$stub:
>
> but can not understand why it compiles using gcc but not when ld is
> call directly. Any idea's would be greatly appreciated.


Try running gcc -v hello.s

That will tell you how gcc is using the linker. You're probably also
missing the crt*.o files (C runtime/startup files). The crt*.o files are
very important if you want to use shared libraries AFAIK, and generally
have the program startup properly. This applies to most unix-like systems
as well. You may not need all of them though. If you just want your start
symbol to be called, and would prefer to not have a .globl main entry
point, you could probably get away with much less.

Also, this is extremely useful when analyzing assembly code:

gcc -S -fverbose-asm file.c

I've been porting a compiler I wrote that generates assembly code. It runs
on another unix-like system now, so I've been spending my spare time
porting it to the Mac. I don't have much spare time though, so things are
slowly progressing. I managed to get some code running, but some syscalls
are failing, because of alignment differences.

A few things I have noticed that are different:

1. there is no .global alias, just .globl
2. there is no .section .rodata, use .cstring
3. C API functions have a _ prefix
4. .equ doesn't seem to work in the Mac assembler (I used a Tcl script for a
sort of scripted assembly instead for my runtime library to work around
some things I couldn't sanely do with cpp macros).
5. sections in general are very different (as your gcc -S output
demonstrated).

I think that's about it.

By the way, there is an official Mac guide to assembly programming online.

-George

On 2009-04-09 00:01:53 +1000, andrew DeleteThis @cucumber.demon.co.uk (Andrew
Gabriel) said:

> In article <49dc9e60$0$12942$afc38c87@news.optusnet.com.au>,
> Aaron Spiteri <azzmosphere(at)gmail.com> writes:
>>
>> All though the program compiled to a executable, When I tried to link
>> it using the following command:
>>
>> as -o hello.o hello.s
>> ld -o hello hello.o
>>
>> I keep getting the message:
>>
>> ld: Undefined symbols:
>> _puts
>>
>> I imagine that this has something to do with the line:
>>
>> L_puts$stub:
>>
>> but can not understand why it compiles using gcc but not when ld is
>> call directly. Any idea's would be greatly appreciated.
>
> Probably because you aren't calling ld the way gcc would.
> For example, you started with C, so you probably need the
> C library for puts(), but ld won't know that.
>
> Try: ld -o hello -lc hello.o
>
> Alternatively, use the gcc command to link it, and it will
> call ld with the correct parameters.

This seems on the right track. I got an executable but when I try
running it I get an illegal instrcution. I think that it may be the
wrong library that I am using.
--
.....
"When asked what he would like on his tomb stone he said more bass"

On 2009-04-09 12:30:30 +1000, George Peter Staplin
<georgeps.DeleteThis@xmission.com> said:

> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>
>> On 2009-04-08 22:41:17 +1000, Aaron Spiteri <azzmosphere(at)gmail.com>
>> said:
>>
>>> On 2009-04-08 04:29:54 +1000, Reinder Verlinde <reinder.DeleteThis@verlinde.invalid>
>>> said:
>>>
>>>> In article <49d9f0a6$0$20975$afc38c87@news.optusnet.com.au>,
>>>> Aaron Spiteri <azzmosphere(at)gmail.com> wrote:
>>>>
>>>>> Does anyone know where I can find some examples of assembly programs in
>>>>> 'as' for intel OSX
>> I wrote a small C program "Hello World" and tested that it compiled.
>> Produced the ASM output and got the following:
>>
>> .cstring
>> LC0:
>> .ascii "Hello World\0"
>> .text
>> .globl _main
>> _main:
>> pushl %ebp
>> movl %esp, %ebp
>> pushl %ebx
>> subl $20, %esp
>> call ___i686.get_pc_thunk.bx
>> "L00000000001$pb":
>> leal LC0-"L00000000001$pb"(%ebx), %eax
>> movl %eax, (%esp)
>> call L_puts$stub
>> movl $0, %eax
>> addl $20, %esp
>> popl %ebx
>> popl %ebp
>> ret
>> .section
>> __IMPORT,__jump_table,symbol_stubs,self_modifying_code+pure_instructions,5
>> L_puts$stub:
>>
>>
>> .indirect_symbol _puts
>> hlt ; hlt ; hlt ; hlt ; hlt
>> .subsections_via_symbols
>> .section __TEXT,__textcoal_nt,coalesced,pure_instructions
>> .weak_definition ___i686.get_pc_thunk.bx
>> .private_extern ___i686.get_pc_thunk.bx
>> ___i686.get_pc_thunk.bx:
>> movl (%esp), %ebx
>> ret
>>
>> All though the program compiled to a executable, When I tried to link
>> it using the following command:
>>
>> as -o hello.o hello.s
>> ld -o hello hello.o
>>
>> I keep getting the message:
>>
>> ld: Undefined symbols:
>> _puts
>>
>> I imagine that this has something to do with the line:
>>
>> L_puts$stub:
>>
>> but can not understand why it compiles using gcc but not when ld is
>> call directly. Any idea's would be greatly appreciated.
>
>
> Try running gcc -v hello.s

I got the following output:
Using built-in specs.
Target: i686-apple-darwin8
Configured with: /private/var/tmp/gcc/gcc-5250.obj~12/src/configure
--disable-checking -enable-werror --prefix=/usr --mandir=/share/man
--enable-languages=c,objc,c++,obj-c++
--program-transform-name=/^[cg][^.-]*$/s/$/-4.0/
--with-gxx-include-dir=/include/c++/4.0.0 --build=powerpc-apple-darwin8
--with-arch=pentium-m --with-tune=prescott --program-prefix=
--host=i686-apple-darwin8 --target=i686-apple-darwin8
Thread model: posix
gcc version 4.0.1 (Apple Computer, Inc. build 5250)
/usr/libexec/gcc/i686-apple-darwin8/4.0.1/cc1 -E -lang-asm -quiet -v
-D__DYNAMIC__ hello.s -fPIC -march=pentium-m -mtune=prescott -o
/var/tmp//cce0s7vu.s
ignoring nonexistent directory
"/usr/lib/gcc/i686-apple-darwin8/4.0.1/../../../../i686-apple-darwin8/include"
#include

"..." search starts here:
#include <...> search starts here:
/usr/local/include
/usr/lib/gcc/i686-apple-darwin8/4.0.1/include
/usr/include
/System/Library/Frameworks
/Library/Frameworks
End of search list.
as -arch i386 -force_cpusubtype_ALL -o /var/tmp//ccEhgz9s.o
/var/tmp//cce0s7vu.s
/usr/libexec/gcc/i686-apple-darwin8/4.0.1/collect2 -dynamic -arch i386
-weak_reference_mismatches non-weak -o a.out -lcrt1.o
/usr/lib/gcc/i686-apple-darwin8/4.0.1/crt2.o
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1/../../.. /var/tmp//ccEhgz9s.o
-lgcc -lSystem

From this the most significant line appears to be

as -arch i386 -force_cpusubtype_ALL -o /var/tmp//ccEhgz9s.o
/var/tmp//cce0s7vu.s
/usr/libexec/gcc/i686-apple-darwin8/4.0.1/collect2 -dynamic -arch i386
-weak_reference_mismatches non-weak -o a.out -lcrt1.o
/usr/lib/gcc/i686-apple-darwin8/4.0.1/crt2.o
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1
-L/usr/lib/gcc/i686-apple-darwin8/4.0.1/../../.. /var/tmp//ccEhgz9s.o
-lgcc -lSystem


>
> That will tell you how gcc is using the linker. You're probably also
> missing the crt*.o files (C runtime/startup files). The crt*.o files are
> very important if you want to use shared libraries AFAIK, and generally
> have the program startup properly. This applies to most unix-like systems
> as well. You may not need all of them though. If you just want your start
> symbol to be called, and would prefer to not have a .globl main entry
> point, you could probably get away with much less.
>
> Also, this is extremely useful when analyzing assembly code:
>
> gcc -S -fverbose-asm file.c
>
> I've been porting a compiler I wrote that generates assembly code. It runs
> on another unix-like system now, so I've been spending my spare time
> porting it to the Mac. I don't have much spare time though, so things are
> slowly progressing. I managed to get some code running, but some syscalls
> are failing, because of alignment differences.
>
> A few things I have noticed that are different:
>
> 1. there is no .global alias, just .globl
> 2. there is no .section .rodata, use .cstring
> 3. C API functions have a _ prefix
> 4. .equ doesn't seem to work in the Mac assembler (I used a Tcl script for a
> sort of scripted assembly instead for my runtime library to work around
> some things I couldn't sanely do with cpp macros).
> 5. sections in general are very different (as your gcc -S output
> demonstrated).
>
> I think that's about it.
>
> By the way, there is an official Mac guide to assembly programming online.
>
> -George


After trying your suggestion I noticed that the a.out file appears that
is the correct executable. I have a fair bit of reading ahead of me to
figure out why this is and exactly what is going on. Thank you very
much for your advice. I found the following URI:

http://developer.apple.com/DOCUMENTATION/DeveloperTools/Reference/Asse...er/010-

Is

this the online programming guide that you where refering too?

--
.....
"When asked what he would like on his tomb stone he said more bass"