Re: [PATCH v5 3.1.0-rc4-tip 26/26] uprobes: queue signals whilethread is singlestepping.

[Oleg Nesterov]

On 10/11, Srikar Dronamraju wrote:
>
> --- a/kernel/uprobes.c
> +++ b/kernel/uprobes.c
> @@ -1366,6 +1366,26 @@ static bool sstep_complete(struct uprobe *uprobe, struct pt_regs *regs)
>  }
>
>  /*
> + * While we are handling breakpoint / singlestep, ensure that a
> + * SIGTRAP is not delivered to the task.
> + */
> +static void __clear_trap_flag(void)
> +{
> +	sigdelset(&current->pending.signal, SIGTRAP);
> +	sigdelset(&current->signal->shared_pending.signal, SIGTRAP);
> +}
> +
> +static void clear_trap_flag(void)
> +{
> +	if (!test_and_clear_thread_flag(TIF_SIGPENDING))
> +		return;
> +
> +	spin_lock_irq(&current->sighand->siglock);
> +	__clear_trap_flag();
> +	spin_unlock_irq(&current->sighand->siglock);
> +}

And this is called before and after the step.

Confused... For what? What makes SIGTRAP special? Where does this
signal come from? If you meant do_debug() this seems impossible,
uprobe_exception_notify(DIE_DEBUG) returns NOTIFY_STOP.

I certainly missed something.

> @@ -1401,13 +1422,18 @@ void uprobe_notify_resume(struct pt_regs *regs)
>  			if (!utask)
>  				goto cleanup_ret;
>  		}
> -		/* TODO Start queueing signals. */
>  		utask->active_uprobe = u;
>  		handler_chain(u, regs);
>  		utask->state = UTASK_SSTEP;
> -		if (!pre_ssout(u, regs, probept))
> +		if (!pre_ssout(u, regs, probept)) {
> +			sigfillset(&masksigs);
> +			sigdelsetmask(&masksigs,
> +					sigmask(SIGKILL)|sigmask(SIGSTOP));
> +			current->saved_sigmask = current->blocked;
> +			set_current_blocked(&masksigs);

OK, we already discussed the problems with this approach.

> +			clear_trap_flag();

In any case unneeded, we already blocked SIGTRAP.

> @@ -1418,8 +1444,8 @@ void uprobe_notify_resume(struct pt_regs *regs)
>  			utask->state = UTASK_RUNNING;
>  			user_disable_single_step(current);
>  			xol_free_insn_slot(current);
> -
> -			/* TODO Stop queueing signals. */
> +			clear_trap_flag();

This is what I can't understand.

> +			set_restore_sigmask();

No, this is not right. If we have a pending signal, the signal handler
will run with the almost-all-blocked mask we set before.

And this is overkill anyway, you could simply do
set_current_blocked(&current->saved_sigmask).

->saved_sigmask is only used when we return from syscall, so uprobes
can (ab)use it safely.

> @@ -1433,7 +1459,7 @@ void uprobe_notify_resume(struct pt_regs *regs)
>  		put_uprobe(u);
>  		set_instruction_pointer(regs, probept);
>  	} else
> -		/*TODO Return SIGTRAP signal */
> +		send_sig(SIGTRAP, current, 0);

This change looks "offtopic" to the problems we are discussing.

Or I missed something and this is connected to the clear_trap_flag()
somehow?

Oleg.

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/