On Wed, Apr 03, 2019 at 08:51:22PM +0300, Leon Romanovsky wrote:
> On Wed, Apr 03, 2019 at 02:38:35PM -0300, Jason Gunthorpe wrote:
> > On Wed, Apr 03, 2019 at 08:15:14PM +0300, Leon Romanovsky wrote:
> > > > @@ -203,28 +263,29 @@ struct ib_umem *ib_umem_get(struct ib_udata *udata, unsigned long addr,
> > > > goto umem_release;
> > > > }
> > > >
> > > > - umem->npages += ret;
> > > > cur_base += ret * PAGE_SIZE;
> > > > npages -= ret;
> > > >
> > > > + sg = ib_umem_add_sg_table(sg, page_list, ret,
> > > > + dma_get_max_seg_size(context->device->dma_device),
> > > > + &umem->sg_nents);
> > > > +
> > > > /* Continue to hold the mmap_sem as vma_list access
> > > > * needs to be protected.
> > > > */
> > > > - for_each_sg(sg_list_start, sg, ret, i) {
> > > > + for (i = 0; i < ret && umem->hugetlb; i++) {
> > > > if (vma_list && !is_vm_hugetlb_page(vma_list[i]))
> > > > umem->hugetlb = 0;
> > > > -
> > > > - sg_set_page(sg, page_list[i], PAGE_SIZE, 0);
> > > > }
> > >
> > > I was under wrong impression that we removed hugetlb flag. Is it still needed?
> > > And more general question, how was 2G limit chosen?
> >
> > The next patches do that - it can't be done until the two drivers
> > using it convert to this new interface.
> >
> > This specific patch is just the only patch in the process that impacts
> > every driver, and is what all the work thus far has been to make safe.
> >
> > Once this is applied it will be easy to apply the next ones as they
> > only impact two drivers and other drivers can debug and use the new
> > features as they need.
>
> Thanks for the explanation, and what about 2G limit?
The sgl->sg_len is an unsigned int so the largest power of two we can
store in it is 2G.
Jason
