linux-kernel - Re: [PATCH v2 2/4] dma-iommu: fix arch_sync

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:   Mon, 2 Aug 2021 14:50:37 +0100
From:   Will Deacon <will@...nel.org>
To:     David Stevens <stevensd@...omium.org>
Cc:     Robin Murphy <robin.murphy@....com>,
        Joerg Roedel <joro@...tes.org>,
        Lu Baolu <baolu.lu@...ux.intel.com>,
        Tom Murphy <murphyt7@....ie>, iommu@...ts.linux-foundation.org,
        linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 2/4] dma-iommu: fix arch_sync_dma for map with swiotlb

On Mon, Aug 02, 2021 at 02:40:59PM +0100, Will Deacon wrote:
> On Fri, Jul 09, 2021 at 12:35:00PM +0900, David Stevens wrote:
> > From: David Stevens <stevensd@...omium.org>
> > 
> > When calling arch_sync_dma, we need to pass it the memory that's
> > actually being used for dma. When using swiotlb bounce buffers, this is
> > the bounce buffer. Move arch_sync_dma into the __iommu_dma_map_swiotlb
> > helper, so it can use the bounce buffer address if necessary. This also
> > means it is no longer necessary to call iommu_dma_sync_sg_for_device in
> > iommu_dma_map_sg for untrusted devices.
> > 
> > Fixes: 82612d66d51d ("iommu: Allow the dma-iommu api to use bounce buffers")
> > Signed-off-by: David Stevens <stevensd@...omium.org>
> > ---
> >  drivers/iommu/dma-iommu.c | 16 +++++++---------
> >  1 file changed, 7 insertions(+), 9 deletions(-)
> > 
> > diff --git a/drivers/iommu/dma-iommu.c b/drivers/iommu/dma-iommu.c
> > index eac65302439e..e79e274d2dc5 100644
> > --- a/drivers/iommu/dma-iommu.c
> > +++ b/drivers/iommu/dma-iommu.c
> > @@ -574,6 +574,9 @@ static dma_addr_t __iommu_dma_map_swiotlb(struct device *dev, phys_addr_t phys,
> >  		memset(padding_start, 0, padding_size);
> >  	}
> >  
> > +	if (!coherent && !(attrs & DMA_ATTR_SKIP_CPU_SYNC))
> > +		arch_sync_dma_for_device(phys, org_size, dir);
> 
> I think this relies on the swiotlb buffers residing in the linear mapping
> (i.e. where phys_to_virt() is reliable), which doesn't look like a safe
> assumption to me.

No, sorry, ignore me here. I misread swiotlb_bounce(), so I think your
change is good.

As an aside, it strikes me that we'd probably be better off using
uncacheable bounce buffers for non-coherent devices so we could avoid all
this maintenance entirely.

Will