linux-kernel - Re: [PATCH v2] iommu/arm-smmu-v3: Match Stall behaviour for S2

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] [day] [month] [year] [list]

Message-ID: <Zr8gom978iZftV_9@google.com>
Date: Fri, 16 Aug 2024 09:49:22 +0000
From: Mostafa Saleh <smostafa@...gle.com>
To: Jason Gunthorpe <jgg@...pe.ca>
Cc: Robin Murphy <robin.murphy@....com>, linux-kernel@...r.kernel.org,
	iommu@...ts.linux.dev, linux-arm-kernel@...ts.infradead.org,
	will@...nel.org, joro@...tes.org, jean-philippe@...aro.org,
	nicolinc@...dia.com, mshavit@...gle.com
Subject: Re: [PATCH v2] iommu/arm-smmu-v3: Match Stall behaviour for S2

On Thu, Aug 15, 2024 at 10:59:51AM -0300, Jason Gunthorpe wrote:
> On Thu, Aug 15, 2024 at 02:41:52PM +0100, Robin Murphy wrote:
> 
> > Unfortunately we can't do that, because there *are* RCiEP devices whose data
> > interfaces are native AMBA, and thus for whom stalling is not actually a
> > protocol violation as it would be on a real PCIe transport layer;
> > correspondingly, it's *because* they are not true PCIe devices that they
> > can't support ATS, and thus need stall support in order to do SVA, so things
> > should still work out OK in practice.
> 
> I wondered if that would be the case.
> 
> Looks like if we want to do anything then arm_smmu_ats_supported()
> would be the right place.

Yes, I guess that might be the place, although the spec seems more
relaxed regarding stage-1 stall.

Anyway I will drop it from this patch, as Robin mentioned this should
only happen with buggy firmware and we can have another patch to
harden that if required.

Thanks,
Mostafa

> 
> > >          if (dev_is_pci(dev)) {
> > > 
> > > Though I have no idea how the GPU driver that wants to use this
> > > works - it doesn't seem to be intree :\
> > 
> > It's not a GPU: drivers/crypto/hisilicon/zip/
> 
> Ohhhh.. so it goes through uacce, I see.
> 
> Jason