Re: vmsplice exploits, stack protector and Makefiles

[pageexec@freemail.hu]

On 14 Feb 2008 at 8:30, Ingo Molnar wrote:

> --- linux-x86.q.orig/arch/x86/kernel/process_64.c
> +++ linux-x86.q/arch/x86/kernel/process_64.c
> @@ -166,6 +166,15 @@ static inline void play_dead(void)
>  void cpu_idle(void)
>  {
>  	current_thread_info()->status |= TS_POLLING;
> +
> +#ifdef CONFIG_CC_STACKPROTECTOR
> +	/*
> +	 * If we're the non-boot CPU, nothing set the PDA stack
> +	 * canary up for us. This is as good a place as any for
> +	 * doing that.
> +	 */
> +	write_pda(stack_canary, current->stack_canary);
> +#endif

i wonder if these #ifdef's are really needed at all, even if one doesn't
use -fstack-protector, having the code set up the canary has like 0
performance impact. not to mention that i think the change in switch_to
means that it won't compile without CONFIG_CC_STACKPROTECTOR enabled and
instead of making that macro conditional it's just a lot easier to enable
the canary all the time.