On 02/05/2014 02:31 AM, Will Deacon wrote:
 On Mon, Feb 03, 2014 at 06:56:30AM +0000, AKASHI Takahiro wrote:
> This patch adds auditing functions on entry to or exit from
> every system call invocation.
>
> Signed-off-by: AKASHI Takahiro <takahiro.akashi(a)linaro.org>
> ---
>   arch/arm64/include/asm/thread_info.h |    1 +
>   arch/arm64/kernel/entry.S            |    3 +++
>   arch/arm64/kernel/ptrace.c           |   10 ++++++++++
>   3 files changed, 14 insertions(+)
>
> diff --git a/arch/arm64/include/asm/thread_info.h
b/arch/arm64/include/asm/thread_info.h
> index 720e70b..7468388 100644
> --- a/arch/arm64/include/asm/thread_info.h
> +++ b/arch/arm64/include/asm/thread_info.h
> @@ -101,6 +101,7 @@ static inline struct thread_info *current_thread_info(void)
>   #define TIF_NEED_RESCHED	1
>   #define TIF_NOTIFY_RESUME	2	/* callback before returning to user */
>   #define TIF_SYSCALL_TRACE	8
> +#define TIF_SYSCALL_AUDIT	9
>   #define TIF_POLLING_NRFLAG	16
>   #define TIF_MEMDIE		18	/* is terminating due to OOM killer */
>   #define TIF_FREEZE		19
> diff --git a/arch/arm64/kernel/entry.S b/arch/arm64/kernel/entry.S
> index 827cbad..83c4b29 100644
> --- a/arch/arm64/kernel/entry.S
> +++ b/arch/arm64/kernel/entry.S
> @@ -630,6 +630,9 @@ el0_svc_naked:					// compat entry point
>   	get_thread_info tsk
>   	ldr	x16, [tsk, #TI_FLAGS]		// check for syscall tracing
>   	tbnz	x16, #TIF_SYSCALL_TRACE, __sys_trace // are we tracing syscalls?
> +#ifdef CONFIG_AUDITSYSCALL
> +	tbnz	x16, #TIF_SYSCALL_AUDIT, __sys_trace // auditing syscalls?
> +#endif
 Could we avoid the back-to-back tbnz instructions with a single mask? It's
 not obvious that it will end up any better, but it would be good to know. 
When first implementing ftrace support, TIF_SYSCALL_TRACEPOINT is defined as 10
and 'tst' instruction doesn't allow the following code:
     tst x16, #(_TIF_SYSCALL_TRACE|_TIF_SYSCALL_TRACEPOINT)
That is why I've used "back-to-back" tbnz since then, but now that I'm
going to
submit ftrace, audit and later seccomp, I will replace it with:
     #define TIF_SYSCALL_WORK (_TIF_SYSCALL_TRACE|TRACEPOINT|AUDIT|SECCOMP)
     tst x16, #TIF_SYSCALL_WORK
     b.ne __syscall_trace
>   	adr	lr, ret_fast_syscall		// return address
>   	cmp     scno, sc_nr                     // check upper syscall limit
>   	b.hs	ni_sys
> diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
> index 6777a21..75a3f23 100644
> --- a/arch/arm64/kernel/ptrace.c
> +++ b/arch/arm64/kernel/ptrace.c
> @@ -19,6 +19,7 @@
>    * along with this program.  If not, see <
http://www.gnu.org/licenses/>.
>    */
>
> +#include <linux/audit.h>
>   #include <linux/kernel.h>
>   #include <linux/sched.h>
>   #include <linux/mm.h>
> @@ -38,6 +39,7 @@
>   #include <asm/compat.h>
>   #include <asm/debug-monitors.h>
>   #include <asm/pgtable.h>
> +#include <asm/syscall.h>
>   #include <asm/traps.h>
>   #include <asm/system_misc.h>
>
> @@ -1064,6 +1066,14 @@ asmlinkage int syscall_trace(int dir, struct pt_regs *regs)
>   {
>   	unsigned long saved_reg;
>
> +	if (dir)
> +		audit_syscall_exit(regs);
> +	else
> +		audit_syscall_entry(syscall_get_arch(current, regs),
> +			(int)regs->syscallno,
> +			regs->orig_x0, regs->regs[1],
> +			regs->regs[2], regs->regs[3]);
> +
 Do we really want to perform the audit checks before the tracehook calls?
 Remember that the latter can rewrite all of the registers. 
OK. I will change the code to make calls in the following order:
On entry,
    *secure_computing
    *tracehook_report_syscall(ENTER)
    *trace_sys_enter
    *audit_syscall_entry
On exit,
    *audit_syscall_exit
    *trace_sys_exit
    *tracehook_report_syscall(EXIT)
The order here is the exact same as on x86, but such change might
decrease the readability in syscall_trace().
Thanks,
-Takahiro AKASHI
 Will