Re: [PATCHv3 33/33] mm, x86: introduce PR_SET_MAX_VADDR and PR_GET_MAX_VADDR

[Kirill A. Shutemov]

On Tue, Feb 21, 2017 at 12:46:55PM -0800, Dave Hansen wrote:
> Let me make sure I'm grokking what you're trying to do here.
> 
> On 02/20/2017 05:15 AM, Kirill A. Shutemov wrote:
> > +/* MPX cannot handle addresses above 47-bits yet. */
> > +unsigned long mpx_unmapped_area_check(unsigned long addr, unsigned long len,
> > +		unsigned long flags)
> > +{
> > +	if (!kernel_managing_mpx_tables(current->mm))
> > +		return addr;
> > +	if (addr + len <= DEFAULT_MAP_WINDOW)
> > +		return addr;
> 
> At this point, we know MPX management is on and the hint is for memory
> above DEFAULT_MAP_WINDOW?

Right.

> > +	if (flags & MAP_FIXED)
> > +		return -ENOMEM;
> 
> ... and if it's a MAP_FIXED request, fail it.

Yep.

> > +	if (len > DEFAULT_MAP_WINDOW)
> > +		return -ENOMEM;
> 
> What is this case for?  If addr+len wraps?

If len is too big to fit into DEFAULT_MAP_WINDOW there's no point in
resetting hint address as we know we can't satisfy it -- fail early.
> 
> > +	/* Look for unmap area within DEFAULT_MAP_WINDOW */
> > +	return 0;
> > +}
> 
> Otherwise, blow away the hint, which we know is high and needs to
> be discarded?

Right.

-- 
 Kirill A. Shutemov