Re: [PATCH v5 3/7] iommu: Allow the dma-iommu api to use bounce buffers
From: Will Deacon
Date: Mon Nov 23 2020 - 06:47:13 EST
On Mon, Nov 23, 2020 at 07:40:57PM +0800, Lu Baolu wrote:
> On 2020/11/23 18:08, Christoph Hellwig wrote:
> > > + /*
> > > + * If both the physical buffer start address and size are
> > > + * page aligned, we don't need to use a bounce page.
> > > + */
> > > + if (IS_ENABLED(CONFIG_SWIOTLB) && dev_is_untrusted(dev) &&
> > > + iova_offset(iovad, phys | org_size)) {
> > > + aligned_size = iova_align(iovad, org_size);
> > > + phys = swiotlb_tbl_map_single(dev,
> > > + phys_to_dma(dev, io_tlb_start),
> > > + phys, org_size, aligned_size, dir, attrs);
> >
> > swiotlb_tbl_map_single takes one less argument in 5.10-rc now.
> >
>
> Yes. But Will's iommu/next branch is based on 5.10-rc3. I synced with
> him, we agreed to keep it 5.10-rc3 and resolve this conflict when
> merging it.
That's right, although I failed to appreciate the conflict was due to a
change in function prototype rather than just a context collision. So
I've updated the vt-d branch to contain the stuff fron Konrad:
https://git.kernel.org/pub/scm/linux/kernel/git/arm64/linux.git/log/?h=for-next/iommu/vt-d
Sorry for messing you around!
Will