Re: [RFC,PATCH 1/2] seccomp_filters: system call filtering usingBPF

[Oleg Nesterov]

On 01/12, Will Drewry wrote:
>
> On Thu, Jan 12, 2012 at 11:23 AM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> > On 01/12, Will Drewry wrote:
> >>
> >> On Thu, Jan 12, 2012 at 10:22 AM, Oleg Nesterov <oleg@xxxxxxxxxx> wrote:
> >> >> +      */
> >> >> +     regs = seccomp_get_regs(regs_tmp, &regs_size);
> >> >
> >> > Stupid question. I am sure you know what are you doing ;) and I know
> >> > nothing about !x86 arches.
> >> >
> >> > But could you explain why it is designed to use user_regs_struct ?
> >> > Why we can't simply use task_pt_regs() and avoid the (costly) regsets?
> >>
> >> So on x86 32, it would work since user_regs_struct == task_pt_regs
> >> (iirc), but on x86-64
> >> and others, that's not true.
> >
> > Yes sure, I meant that userpace should use pt_regs too.
> >
> >> If it would be appropriate to expose pt_regs to userspace, then I'd
> >> happily do so :)
> >
> > Ah, so that was the reason. But it is already exported? At least I see
> > the "#ifndef __KERNEL__" definition in arch/x86/include/asm/ptrace.h.
> >
> > Once again, I am not arguing, just trying to understand. And I do not
> > know if this definition is part of abi.
>
> I don't either :/  My original idea was to operate on task_pt_regs(current),
> but I noticed that PTRACE_GETREGS/SETREGS only uses the
> user_regs_struct. So I went that route.

Well, I don't know where user_regs_struct come from initially. But
probably it is needed to allow to access the "artificial" things like
fs_base. Or perhaps this struct mimics the layout in the coredump.

> I'd love for pt_regs to be fair game to cut down on the copying!

Me too. I see no point in using user_regs_struct.

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/