Re: [patch -mm][Intel-IOMMU] Optimize sg map/unmap calls

["Keshavamurthy, Anil S" <anil.s.keshavamurthy@intel.com>]

On Wed, Aug 01, 2007 at 04:45:54PM -0700, Andrew Morton wrote:
> On Wed, 1 Aug 2007 13:06:23 -0700
> "Keshavamurthy, Anil S" <anil.s.keshavamurthy@intel.com> wrote:
> 
> > +/* Computes the padding size required, to make the
> > + * the start address naturally aligned on its size
> > + */
> > +static int
> > +iova_get_pad_size(int size, unsigned int limit_pfn)
> > +{
> > +	unsigned int pad_size = 0;
> > +	unsigned int order = ilog2(size);
> > +
> > +	if (order)
> > +		pad_size = (limit_pfn + 1) % (1 << order);
> > +
> > +	return pad_size;
> > +}
> 
> This isn't obviously doing the right thing for non-power-of-2 inputs.
> ilog2() rounds down...
Andrew,
	The call chain to iova_get_pad_size() is like this
alloc_iova()--->__alloc_iova_range()--->iova_get_pad_size().

Inside the alloc_iova() we are rounding the size to __roundup_pow_of_two(size)
iff the caller of alloc_iova() request by setting size_aligned bool. And
in every call to iova_get_pad_size() we check the size_aligned bool before calling
iova_get_pad_size.  Hence I don;t see any issues. If you want I can insert a 
BUG_ON() statement inside the above iova_get_pad_size() function. 
Please do let me know.

thanks,
Anil