boringssl/crypto/perlasm
David Benjamin 17d553d299 Add a CFI tester to CHECK_ABI.
This uses the x86 trap flag and libunwind to test CFI works at each
instruction. For now, it just uses the system one out of pkg-config and
disables unwind tests if unavailable. We'll probably want to stick a
copy into //third_party and perhaps try the LLVM one later.

This tester caught two bugs in P-256 CFI annotations already:
I47b5f9798b3bcee1748e537b21c173d312a14b42 and
I9f576d868850312d6c14d1386f8fbfa85021b347

An earlier design used PTRACE_SINGLESTEP with libunwind's remote
unwinding features. ptrace is a mess around stop signals (see group-stop
discussion in ptrace(2)) and this is 10x faster, so I went with it. The
question of which is more future-proof is complex:

- There are two libunwinds with the same API,
  https://www.nongnu.org/libunwind/ and LLVM's. This currently uses the
  system nongnu.org for convenience. In future, LLVM's should be easier
  to bundle (less complex build) and appears to even support Windows,
  but I haven't tested this.  Moreover, setting the trap flag keeps the
  test single-process, which is less complex on Windows. That suggests
  the trap flag design and switching to LLVM later. However...

- Not all architectures have a trap flag settable by userspace. As far
  as I can tell, ARMv8's PSTATE.SS can only be set from the kernel. If
  we stick with nongnu.org libunwind, we can use PTRACE_SINGLESTEP and
  remote unwinding. Or we implement it for LLVM. Another thought is for
  the ptracer to bounce SIGTRAP back into the process, to share the
  local unwinding code.

- ARMv7 has no trap flag at all and PTRACE_SINGLESTEP fails. Debuggers
  single-step by injecting breakpoints instead. However, ARMv8's trap
  flag seems to work in both AArch32 and AArch64 modes, so we may be
  able to condition it on a 64-bit kernel.

Sadly, neither strategy works with Intel SDE. Adding flags to cpucap
vectors as we do with ARM would help, but it would not emulate CPUs
newer than the host CPU. For now, I've just had SDE tests disable these.

Annoyingly, CMake does not allow object libraries to have dependencies,
so make test_support a proper static library. Rename the target to
test_support_lib to avoid
https://gitlab.kitware.com/cmake/cmake/issues/17785

Update-Note: This adds a new optional test dependency, but it's disabled
by default (define BORINGSSL_HAVE_LIBUNWIND), so consumers do not need
to do anything. We'll probably want to adjust this in the future.

Bug: 181
Change-Id: I817263d7907aff0904a9cee83f8b26747262cc0c
Reviewed-on: https://boringssl-review.googlesource.com/c/33966
Commit-Queue: David Benjamin <davidben@google.com>
Reviewed-by: Adam Langley <agl@google.com>
2019-01-03 22:01:55 +00:00
..
arm-xlate.pl Add a note that generated files are generated. 2018-11-21 20:05:05 +00:00
ppc-xlate.pl Add a note that generated files are generated. 2018-11-21 20:05:05 +00:00
readme
x86_64-xlate.pl Add a CFI tester to CHECK_ABI. 2019-01-03 22:01:55 +00:00
x86asm.pl Make symbol-prefixing work on 32-bit x86. 2018-11-27 22:35:17 +00:00
x86gas.pl
x86masm.pl
x86nasm.pl

The perl scripts in this directory are my 'hack' to generate
multiple different assembler formats via the one origional script.

The way to use this library is to start with adding the path to this directory
and then include it.

push(@INC,"perlasm","../../perlasm");
require "x86asm.pl";

The first thing we do is setup the file and type of assembler

&asm_init($ARGV[0]);

The first argument is the 'type'.  Currently
'cpp', 'sol', 'a.out', 'elf' or 'win32'.
Argument 2 is the file name.

The reciprocal function is
&asm_finish() which should be called at the end.

There are 2 main 'packages'. x86ms.pl, which is the Microsoft assembler,
and x86unix.pl which is the unix (gas) version.

Functions of interest are:
&external_label("des_SPtrans");	declare and external variable
&LB(reg);			Low byte for a register
&HB(reg);			High byte for a register
&BP(off,base,index,scale)	Byte pointer addressing
&DWP(off,base,index,scale)	Word pointer addressing
&stack_push(num)		Basically a 'sub esp, num*4' with extra
&stack_pop(num)			inverse of stack_push
&function_begin(name,extra)	Start a function with pushing of
				edi, esi, ebx and ebp.  extra is extra win32
				external info that may be required.
&function_begin_B(name,extra)	Same as normal function_begin but no pushing.
&function_end(name)		Call at end of function.
&function_end_A(name)		Standard pop and ret, for use inside functions
&function_end_B(name)		Call at end but with poping or 'ret'.
&swtmp(num)			Address on stack temp word.
&wparam(num)			Parameter number num, that was push
				in C convention.  This all works over pushes
				and pops.
&comment("hello there")		Put in a comment.
&label("loop")			Refer to a label, normally a jmp target.
&set_label("loop")		Set a label at this point.
&data_word(word)		Put in a word of data.

So how does this all hold together?  Given

int calc(int len, int *data)
	{
	int i,j=0;

	for (i=0; i<len; i++)
		{
		j+=other(data[i]);
		}
	}

So a very simple version of this function could be coded as

	push(@INC,"perlasm","../../perlasm");
	require "x86asm.pl";
	
	&asm_init($ARGV[0]);

	&external_label("other");

	$tmp1=	"eax";
	$j=	"edi";
	$data=	"esi";
	$i=	"ebp";

	&comment("a simple function");
	&function_begin("calc");
	&mov(	$data,		&wparam(1)); # data
	&xor(	$j,		$j);
	&xor(	$i,		$i);

	&set_label("loop");
	&cmp(	$i,		&wparam(0));
	&jge(	&label("end"));

	&mov(	$tmp1,		&DWP(0,$data,$i,4));
	&push(	$tmp1);
	&call(	"other");
	&add(	$j,		"eax");
	&pop(	$tmp1);
	&inc(	$i);
	&jmp(	&label("loop"));

	&set_label("end");
	&mov(	"eax",		$j);

	&function_end("calc");

	&asm_finish();

The above example is very very unoptimised but gives an idea of how
things work.