Re: [PATCH v2] intel-iommu: Quiesce devices before disabling IOMMU

From: David Woodhouse
Date: Tue Sep 24 2013 - 09:39:44 EST

On Tue, 2013-09-24 at 15:16 +0200, Joerg Roedel wrote:
> I am not convinced that this is the right approach. If a device wasn't
> translated by VT-d in the old kernel doesn't mean it will not be
> translated in the new kernel. How about unconditionally resetting all
> PCI busses and/or functions here before IOMMU initialization proceeds?

Forget the IOMMU.

If a device is doing DMA in the old kernel, and *continues* doing DMA
under the new kernel, surely something ought to shut it down?

It's the generic kexec or machine_shutdown code which should be doing
this, not IOMMU code. *Especially* not Intel-specific IOMMU code. It
does not live here.

If anything, the IOMMU gives you a way to *survive* this kind of thing
and means that you might get away without quiescing devices when
otherwise you would have died.


Attachment: smime.p7s
Description: S/MIME cryptographic signature