linux-kernel - Re: [PATCH 1/2] dma-direct: Validate DMA mask against canonical DMA addresses

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]

Message-ID: <20260120195428.GW961572@ziepe.ca>
Date: Tue, 20 Jan 2026 15:54:28 -0400
From: Jason Gunthorpe <jgg@...pe.ca>
To: Robin Murphy <robin.murphy@....com>
Cc: Suzuki K Poulose <suzuki.poulose@....com>,
	"Aneesh Kumar K.V" <aneesh.kumar@...nel.org>,
	linux-kernel@...r.kernel.org, iommu@...ts.linux.dev,
	linux-coco@...ts.linux.dev,
	Catalin Marinas <catalin.marinas@....com>, will@...nel.org,
	steven.price@....com, Marek Szyprowski <m.szyprowski@...sung.com>
Subject: Re: [PATCH 1/2] dma-direct: Validate DMA mask against canonical DMA
 addresses

On Tue, Jan 20, 2026 at 06:47:02PM +0000, Robin Murphy wrote:
> > > then the fact is still that DA requires an SMMU for S2, so at worst
> > > there should always be the possibility for an RMM to offer S1 SMMU
> > > support to the Realm, we're just not there yet.
> > 
> > Having a S1 would help a T=1 device, but it doesn't do anything for
> > the T=0 devices.
> 
> If we have an SMMU we have an SMMU - S1 for T=0 devices is just regular
> VFIO/IOMMUFD in Non-Secure VA/IPA space, for which the VMM doesn't need the
> RMM's help. I've long been taking it for granted that that one's a non-issue
> ;)

Well, sort of. Yes that's all true, but since the T=0 vSMMU cannot
access private memory it will not work properly with the current
driver in Linux which doesn't know how to allocate shared memory for
the page table.

> The only thing we can't easily handle (and would rather avoid) is S1
> translation for T=0 traffic from T=1 devices, since that would require the
> Realm OS to comprehend the notion of a single device attached to two
> different vSMMUs at once.

Yes, so far the general consensus I've heard is that this will not be
done, the vSMMU presented to the VM will only handle T=1 traffic and
the OS must assume that T=0 traffic is identity.

Given those two reasons my general take is that we will not see a T=0
vSMMU in ccVMs at all.

>  Rather, to be workable I think we'd need to keep the T=0 and T=1
> states described as distinct devices - which *could* then each be
> associated with "shared" (VMM-provided) and "private" (RMM-provided)
> vSMMU instances respectively - and leave it as the Realm driver's
> problem if it wants to coordinate enabling and using both at the
> same time, kinda like the link aggregation model.

Hopefully we don't need to do this :(

The current thought is that there would be only one device and when it
is in T=0 mode the OS knows the linked iommu is not being used on
ARM. IIRC AMD and Intel do not have this quirk.

Sadly there are use cases for a TDISP device to do T=0 DMA before it
reaches the run state.

Jason