invalid instruction suffix for mov? - gcc

I have this piece of inline assembly code that should print A in text mode:
void print(){
asm volatile(
"mov ax,0xb800\n"
"mov ds,ax\n" /*<-as complains about this*/
"movb 0,'A'\n"
);
}
However when I try to compile it with gcc(with -m32 and -masm=intel):
./source/kernel.c: Assembler messages:
./source/kernel.c:4: Error: invalid instruction suffix for `mov'
btw this piece of code is from my operating system's kernel, so I can't use stdio.h or something like that.

Despite GCC's line numbering in the error message, that's not the line it's actually complaining about, it's the movb store. You can test that by commenting the other instructions. The error is actually printed by the assembler, with numbering based on .loc metadata directives from the compiler, and this is a multi-line asm template, so it's easy for that to go wrong I guess.
I suspect GAS .intel_syntax mode treats a literal 0 as an immediate, for consistency with mov al, 0 with 0 as a source operand. This of course can't work as a destination.
The "invalid instruction suffix" error message makes little sense, although note that Intel syntax doesn't use operand-size suffixes. (For some reason movb [0], 'A' is accepted, though.)
Instead use square brackets to avoid ambiguity; recommended for any memory operand, even if the address is a symbol instead of a literal number.
mov byte ptr [0], 'A'
mov byte ptr ds:0, 'A' also works, and is the syntax objdump -d -Mintel uses.
It's a good idea to always use square brackets on memory operands to remove any ambiguity, especially for people who might be used to the NASM flavour of Intel syntax.

Related

How can I get a reference to standard input from assembly on a mac?

I'm following along in some books on assembler and find that they happily use fprintf with stdout. They simply reference stdout as a known symbol. I tried to do that in my code, and the linker complains that stdout is not found. I tried variation on spelling including a leading underscore. I also disassembled printf. It sets up a call to vfprintf using the following line:
leaq 0x41879948(%rip), %rax ; __stdoutp
So it seems like I should be able to do something like:
leaq __stdoutp(%rip), %rdi
Didn't work. Linker complains of undefined symbol. Tried with variations. Left off the p suffix, one or both of the underscores. Nothing worked.
Any ideas, or insider knowledge I can follow to access to this symbol?
Based on #PeterCordes suggestion of looking at the compiler generated asm, I wrote a c program:
#include <stdio.h>
int main() { fprintf(stdout, "Hello, world!\n"); }
ran clang -S main.c on it, and had a look at the asm. Here's how clang generated asm is accessing the stdout global
movq ___stdoutp#GOTPCREL(%rip), %rax
movq (%rax), %rdi
Turns out this is stated in the ABI (section 3.5.4)
Position-independent code cannot contain absolute address. To access a global symbol the address of the symbol has to be loaded from the Global Offset Table. The address of
the entry in the GOT can be obtained with a %rip-relative instruction in the small model.
stdout in the c code refers to the __stdoutp symbol. This becomes ___stdoutp in assembler, because macosx. Similar to how main is _main.
ABI section 5.2 explains GOT this way
Global offset tables hold absolute addresses in private data, thus making the addresses available without compromising the position-independence and shareability of a program’s text.

JS instruction - Hex code

I tried to cover asm to hex by gcc ( gcc -c -masm=intel code.s) by it error: code.s:1: Error: too many memory references for `mov'.
code.s asm:
mov rax , 0x00000000004007E6
js rax
what happened? Can you tutorial me cover it?
As Margaret Bloom said, -masm=intel is only a code generation option, meaning it affects what kind of assembly the compiler generates from C code. It does not affect assembling at all.
You need to put .intel_syntax noprefix directive in your assembly file to switch syntax, or use the convoluted gcc -Wa,--msyntax=intel,-mnaked-reg method to pass argument to the assembler.
Having done the above, the next issue will be that js rax is invalid as there is no conditional indirect jump in x86. The easiest will be probably to use a reversed condition, jns, to skip over a jmp rax.

How to set gcc or clang to use Intel syntax permanently for inline asm() statements?

I have the following code which compiles fine with the gcc command gcc ./example.c. The program itself calls the function "add_two" which simply adds two integers. To use the intel syntax within the extended assembly instructions I need to switch at first to intel and than back to AT&T. According to the gcc documentation it is possible to switch to intel syntax entirely by using gcc -masm=intel ./exmaple.
Whenever I try to compile it with the switch -masm=intel it won't compile and I don't understand why? I already tried to delete the instruction .intel_syntax but it still don't compile.
#include <stdio.h>
int add_two(int, int);
int main(){
int src = 3;
int dst = 5;
printf("summe = %d \n", add_two(src, dst));
return 0;
}
int add_two(int src, int dst){
int sum;
asm (
".intel_syntax;" //switch to intel syntax
"mov %0, %1;"
"add %0, %2;"
".att_syntax;" //switch to at&t syntax
: "=r" (sum) //output
: "r" (src), "r" (dst) //input
);
return sum;
}
The error message by compiling the above mentioned program with gcc -masm=intel ./example.c is:
tmp/ccEQGI4U.s: Assembler messages:
/tmp/ccEQGI4U.s:55: Error: junk `PTR [rbp-4]' after expression
/tmp/ccEQGI4U.s:55: Error: too many memory references for `mov'
/tmp/ccEQGI4U.s:56: Error: too many memory references for `mov'
Use -masm=intel and don't use any .att_syntax directives in your inline asm. This works with GCC and I think ICC, and with any constraints you use. Other methods don't. (See Can I use Intel syntax of x86 assembly with GCC? for a simple answer saying that; this answer explores exactly what goes wrong, including with clang 13 and earlier.)
That also works in clang 14 and later. (Which isn't released yet but the patch is part of current trunk; see https://reviews.llvm.org/D113707).
Clang 13 and earlier would always use AT&T syntax for inline asm, both in substituting operands and in assembling as op src, dst. But even worse, clang -masm=intel would do that even when taking the Intel side of an asm template using dialect-alternatives like asm ("add {att | intel}" : ... )`!
clang -masm=intel did still control how it printed asm after its built-in assembler turned an asm() statement into some internal representation of the instruction. e.g. Godbolt showing clang13 -masm=intel turning add %0, 1 as add dword ptr [1], eax, but clang trunk producing add eax, 1.
Some of the rest of this answer talking about clang hasn't been updated for this new clang patch.
Clang does support Intel-syntax inside MSVC-style asm-blocks, but that's terrible (no constraints so inputs / outputs have to go through memory.
If you were hard-coding register names with clang, -masm=intel would be usable (or the equivalent -mllvm --x86-asm-syntax=intel). But it chokes on mov %eax, 5 in Intel-syntax mode so you can't let %0 expand to an AT&T-syntax register name.
-masm=intel makes the compiler use .intel_syntax noprefix at the top of its asm output file, and use Intel-syntax when generating asm from C outside your inline-asm statement. Using .att_syntax at the bottom of your asm template breaks the compiler's asm, hence the error messages like PTR [rbp-4] looking like junk to the assembler (which is expecting AT&T syntax).
The "too many operands for mov" is because in AT&T syntax, mov eax, ebx is a mov from a memory operand (with symbol name eax) to a memory operand (with symbol name ebx)
Some people suggest using .intel_syntax noprefix and .att_syntax prefix around your asm template. That can sometimes work but it's problematic. And incompatible with the preferred method of -masm=intel.
Problems with the "sandwich" method:
When the compiler substitutes operands into your asm template, it will do so according to -masm=. This will always break for memory operands (the addressing-mode syntax is completely different).
It will also break with clang even for registers. Clang's built-in assembler does not accept %eax as a register name in Intel-syntax mode, and doesn't accept .intel_syntax prefix (as opposed to the noprefix that's usually used with Intel-syntax).
Consider this function:
int foo(int x) {
asm(".intel_syntax noprefix \n\t"
"add %0, 1 \n\t"
".att_syntax"
: "+r"(x)
);
return x;
}
It assembles as follows with GCC (Godbolt):
movl %edi, %eax
.intel_syntax noprefix
add %eax, 1 # AT&T register name in Intel syntax
.att_syntax
The sandwich method depends on GAS accepting %eax as a register name even in Intel-syntax mode. GAS from GNU Binutils does, but clang's built-in assembler doesn't.
On a Mac, even using real GCC the asm output has to assemble with an as that's based on clang, not GNU Binutils.
Using clang on that source code complains:
<source>:2:35: error: unknown token in expression
asm(".intel_syntax noprefix \n\t"
^
<inline asm>:2:6: note: instantiated into assembly here
add %eax, 1
^
(The first line of the error message didn't handle the multi-line string literal very well. If you use ; instead of \n\t and put everything on one line the clang error message works better but the source is a mess.)
I didn't check what happens with "ri" constraints when the compiler picks an immediate; it will still decorate it with $ but IDK if GAS silently ignores that, too, in Intel syntax mode.
PS: your asm statement has a bug: you forgot an early-clobber on your output operand so nothing is stopping the compiler from picking the same register for the %0 output and the %2 input that you don't read until the 2nd instruction. Then mov will destroy an input.
But using mov as the first or last instruction of an asm-template is usually also a missed-optimization bug. In this case you can and should just use lea %0, [%1 + %2] to let the compiler add with the result written to a 3rd register, non-destructively. Or just wrap the add instruction (using a "+r" operand and an "r", and let the compiler worry about data movement.) If it had to load the value from memory anyway, it can put it in the right register so no mov is needed.
PS: it's possible to write inline asm that works with -masm=intel or att, using GNU C inline asm dialect alternatives. e.g.
void atomic_inc(int *p) {
asm( "lock add{l $1, %0 | %0, 1}"
: "+m" (*p)
:: "memory"
);
}
compiles with gcc -O2 (-masm=att is the default) to
atomic_inc(int*):
lock addl $1, (%rdi)
ret
Or with -masm=intel to:
atomic_inc(int*):
lock add DWORD PTR [rdi], 1
ret
Notice that the l suffix is required for AT&T, and the dword ptr is required for intel, because memory, immediate doesn't imply an operand-size. And that the compiler filled in valid addressing-mode syntax for both cases.
This works with clang, but only the AT&T version ever gets used.
Note that -masm= also affects the default inline assembler syntax:
Output assembly instructions using selected dialect. Also affects
which dialect is used for basic "asm" and extended "asm". Supported
choices (in dialect order) are att or intel. The default is att.
Darwin does not support intel.
That means that your first .intel_syntax directive is superfluous and the final .att_syntax is wrong because your GCC call compiles C to Intel assembler code.
IOW, either stick to -masm=intel or sandwich your inline Intel assembler code sections between .intel_syntax noprefix and .att_syntax prefix directives - but don't do both.
Note that the sandwich method isn't compatible with all inline assembler constraints - e.g. a constraint that involves m (i.e. memory operand) would insert an operand in ATT syntax which would yield an error like 'Error: junk (%rbp) after expression'. In those cases you have to use -masm=intel.

Assembly print variables and values

I have this code
global start
section .text
start:
mov rax,0x2000004
mov rdi,1
mov rsi,msg
mov rdx,msg.len
syscall
mov rax,0x2000004
mov rdi,2
mov rsi,msgt
mov rdx,msgt.len
syscall
mov rax,0x2000004
mov rdi,3
mov rsi,msgtn
mov rdx,msgtn.len
syscall
mov rax,0x2000001
mov rdi,0
syscall
section .data
msg: db "This is a string",10
.len: equ $ - msg
var: db 1
msgt: db "output of 1+1: "
.len: equ $ - msgt
msgtn: db 1
.len: equ $ - msg
I want to print the variable msgtn. I tried msgt: db "output of 1+1", var
But the NASM assembler failed with:
second.s:35: error: Mach-O 64-bit format does not support 32-bit absolute addresses
Instead of the variable, I also tried "output of 1+1", [1+1], but I got:
second.s:35: error: expression syntax error
I tried it also without the parantheses, there was no number, but only the string "1+1".
The command I used to assemble my program was:
/usr/local/Cellar/nasm/*/bin/nasm -f macho64 second.s && ld -macosx_version_min 10.7.0 second.o second.o
nasm -v shows:
NASM version 2.11.08 compiled on Nov 27 2015
OS X 10.9.5 with Intel core i5 (x86_64 assembly)
db directives let you put assemble-time-constant bytes into the object file (usually in the data section). You can use an expression as an argument, to have the assembler do some math for you at assemble time. Anything that needs to happen at run time needs to be done by instructions that you write, and that get run. It's not like C++ where a global variable can have a constructor that gets run at startup behind the scenes.
msgt: db "output of 1+1", var
would place those ascii characters, followed by (the low byte of?) the absolute address of var. You'd use this kind of thing (with dd or dq) to do something like this C: int var; int *global_ptr = &var;, where you have a global/static pointer variable that starts out initialized to point to another global/static variable. I'm not sure if MacOS X allows this with a 64bit pointer, or if it just refuses to do relocations for 32bit addresses. But that's why you're getting:
second.s:35: error: Mach-O 64-bit format does not support 32-bit absolute addresses
Notice that numeric value of the pointer depends on where in virtual address space the code is loaded. So the address isn't strictly an assemble-time constant. The linker needs to mark things that need run-time relocation, like those 64bit immediate-constant addresses you mov into registers (mov rsi,msg). See this answer for some information on the difference between that and lea rsi, [rel msg] to get the address into a register using a RIP-relative method. (That answer has links to more detailed info, and so do the x86 wiki).
Your attempt at using db [1+1]: What the heck were you expecting? [] in NASM syntax means memory reference. First: the resulting byte has to be an assemble-time constant. I'm not sure if there's an easy syntax for duplicating whatever's at some other address, but this isn't it. (I'd just define a macro and use it in both places.) Second: 2 is not a valid address.
msgt: db "output of 1+1: ", '0' + 1 + 1, 10
would put the ASCII characters: output of 1+1: 2\n at that point in the object file. 10 is the decimal value of ASCII newline. '0' is a way of writing 0x30, the ASCII encoding the character '0'. A 2 byte is not a printable ASCII character. Your version that did that would have printed a 2 byte there, but you wouldn't notice unless you piped the output into hexdump (or od -t x1c or something, IDK what OS X provides. od isn't very nice, but it is widely available.)
Note that this string is not null-terminated. If you want to pass it to something expecting an implicit-length string (like fputs(3) or strchr(3), instead of write(2) or memchr(3)), tack on an extra , 0 to add a zero-byte after everything else.
If you wanted to do the math at run-time, you need to get data into register, add it, then store a string representation of the number into a buffer somewhere. (Or print it one byte at a time, but that's horrible.)
The easy way is to just call printf, to easily print a constant string with some stuff substituted in. Spend your time writing asm for the part of your code that needs to be hand-tuned, not re-implementing library functions.
There's some discussion of int-to-string in comments.
Your link command looks funny:
ld -macosx_version_min 10.7.0 second.o second.o
Are you sure you want the same .o twice?
You could save some code bytes by only moving to 32bit registers when you don't need sign-extension into the 64bit reg. e.g. mov edi,2 instead of mov rdi,2 saves a byte (the REX prefix), unless NASM is clever and does that anyway (actually, it does).
lea rsi, [rel msg] (or use default rel) is a shorter instruction than mov r64, imm64, though. (The AT&T mnemonic is movabs, but Intel syntax still calls it mov.)

Intel assembly syntax OFFSET

Now that i know u can use gcc for Intel syntax instead of default at&t with
gcc -S -masm=intel test.c
There is this line
mov DWORD PTR [ebp-16], OFFSET FLAT:base
Is it the same as mov dword[ebp-16], base?
Otherwise what must i do?
Yes, mov dword [ebp - 16], base is correct NASM syntax to store the label address to 4 bytes of memory at EBP-16.
I haven't seen offset flat: for a while - I think it's obsolete, but it's what GAS's idea of .intel_syntax noprefix used to demand (I had to look at Gas's source code to find that out). gcc -masm=intel uses it when using symbol addresses as immediates, but offset base works, too, in GAS.
It means the same as offset to MASM, or the unadorned variable name in NASM.

Resources