Re: [PATCH 6/7] uprobes: Do not allocate current->utask unnecessary

[Oleg Nesterov]

On 01/08, Srikar Dronamraju wrote:
>
> * Oleg Nesterov <oleg@xxxxxxxxxx> [2012-12-31 18:52:29]:
>
> >  static void handle_swbp(struct pt_regs *regs)
> >  {
> > -	struct uprobe_task *utask;
> >  	struct uprobe *uprobe;
> >  	unsigned long bp_vaddr;
> >  	int uninitialized_var(is_swbp);
> > @@ -1512,19 +1515,12 @@ static void handle_swbp(struct pt_regs *regs)
> >  	if (unlikely(!test_bit(UPROBE_COPY_INSN, &uprobe->flags)))
> >  		goto out;
> >
> > -	utask = get_utask();
> > -	if (!utask)
> > -		goto out;	/* re-execute the instruction. */
> > -
>
> If get_utask fails with the above change, Dont we end up calling
> handler_chain twice(or more)?.

After restart, yes.

> I think this is probably true with
> previous patch too.

And this can happen with the current code too, if xol_alloc_area()
fails. So I think this is probably fine. Besides, if GFP_KERNEL
fails the task should be oom-killed in practice.

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/