Re: [PATCH v7 1/6] arm64: ptrace: add PTRACE_SET_SYSCALL

[Will Deacon <will.deacon@arm.com>]

Hi Akashi,

On Thu, Oct 02, 2014 at 10:46:11AM +0100, AKASHI Takahiro wrote:
> To allow tracer to be able to change/skip a system call by re-writing
> a syscall number, there are several approaches:
> 
> (1) modify x8 register with ptrace(PTRACE_SETREGSET), and handle this case
>     later on in syscall_trace_enter(), or
> (2) support ptrace(PTRACE_SET_SYSCALL) as on arm
> 
> Thinking of the fact that user_pt_regs doesn't expose 'syscallno' to
> tracer as well as that secure_computing() expects a changed syscall number
> to be visible, especially case of -1, before this function returns in
> syscall_trace_enter(), we'd better take (2).
> 
> Reviewed-by: Kees Cook <keescook@chromium.org>
> Signed-off-by: AKASHI Takahiro <takahiro.akashi@linaro.org>
> ---
>  arch/arm64/include/uapi/asm/ptrace.h |    1 +
>  arch/arm64/kernel/ptrace.c           |   14 +++++++++++++-
>  2 files changed, 14 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/include/uapi/asm/ptrace.h b/arch/arm64/include/uapi/asm/ptrace.h
> index 6913643..49c6174 100644
> --- a/arch/arm64/include/uapi/asm/ptrace.h
> +++ b/arch/arm64/include/uapi/asm/ptrace.h
> @@ -23,6 +23,7 @@
>  
>  #include <asm/hwcap.h>
>  
> +#define PTRACE_SET_SYSCALL	23
>  
>  /*
>   * PSR bits
> diff --git a/arch/arm64/kernel/ptrace.c b/arch/arm64/kernel/ptrace.c
> index fe63ac5..2842f9f 100644
> --- a/arch/arm64/kernel/ptrace.c
> +++ b/arch/arm64/kernel/ptrace.c
> @@ -1082,7 +1082,19 @@ const struct user_regset_view *task_user_regset_view(struct task_struct *task)
>  long arch_ptrace(struct task_struct *child, long request,
>  		 unsigned long addr, unsigned long data)
>  {
> -	return ptrace_request(child, request, addr, data);
> +	int ret;
> +
> +	switch (request) {
> +		case PTRACE_SET_SYSCALL:
> +			task_pt_regs(child)->syscallno = data;
> +			ret = 0;
> +			break;
> +		default:
> +			ret = ptrace_request(child, request, addr, data);
> +			break;
> +	}
> +
> +	return ret;
>  }

I still don't understand why this needs to be in arch-specific code. Can't
we implement this in generic code and get architectures to implement
something like syscall_set_nr if they want the generic interface?

Will