Re: [PATCH 1/2] fork: add clone6

[Andy Lutomirski]

On Tue, May 28, 2019 at 3:08 AM Christian Brauner <christian@xxxxxxxxxx> wrote:
>
> On Mon, May 27, 2019 at 12:27:08PM -0700, Linus Torvalds wrote:
> > On Mon, May 27, 2019 at 3:42 AM Christian Brauner <christian@xxxxxxxxxx> wrote:
> > >
> > > Hm, still pondering whether having one unsigned int argument passed
> > > through registers that captures all the flags from the old clone() would
> > > be a good idea.
> >
> > That sounds like a reasonable thing to do.
> >
> > Maybe we could continue to call the old flags CLONE_XYZ and continue
> > to pass them in as "flags" argument, and then we have CLONE_EXT_XYZ
> > flags for a new 64-bit flag field that comes in through memory in the
> > new clone_args thing?
>
> Hm. I think I'll try a first version without an additional register
> flags argument. And here's why: I'm not sure it buys us a lot especially
> if we're giving up on making this convenient for seccomp anyway.
> And with that out of the way (at least for the moment) I would really
> like to make this interface consistent. But we can revisit this when I
> have the code.
>

Seems reasonable.  Once the interface is nailed down, we can see if it
makes sense to break out some flags into a register.  I would guess
that all the unsharing flags are a good candidate.