On Tue, Jun 09, 2015 at 10:53:23AM +0100, Chris Wilson wrote:
> > > if (!i || page_to_pfn(page) != last_pfn + 1) {
> > > if (i)
> > > - sg = sg_next(sg);
> > > + sg = sg_next(end = sg);
> >
> > But this looks incorrect to me, sg points now to the last element for
> > which we set the page below and end points to one before the last. Or
> > I'm (still) missing something..
>
> Nah, this was a change I made to try and reduce te patch size.
> Originally I did sg_set_page(); end = sg; but then thought I could avoid
> resetting end to the same sg in a few cases.
And without the complication of swiotlb_nr_tlb(), then sg after the loop
is pointing to the last set page anyway. The current code is correct,
sorry for the noise.
-Chris
--
Chris Wilson, Intel Open Source Technology Centre
_______________________________________________
Intel-gfx mailing list
Intel-gfx@xxxxxxxxxxxxxxxxxxxxx
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
