Re: [Qemu-devel] issues of region cache and iommu reset

[Michael S. Tsirkin]

On Wed, Mar 29, 2017 at 05:24:18PM +0800, Jason Wang wrote:
> 
> 
> On 2017年03月29日 17:11, Paolo Bonzini wrote:
> > 
> > On 29/03/2017 11:09, Jason Wang wrote:
> > > > > This looks a good fix but may not solve this issue completely. Depends
> > > > > on the iova that guest uses, address_space_cache_init() may succeed 
> > > > > even
> > > > > in this case since vtd does a passthrough translation in this case.
> > > > Can you explain this more?
> > > Yes.  In this case, virtio-net-pci is not reset, but vtd has been reset.
> > > So virtio-net-pci will still try to use the iova of ring to setup the
> > > cache but now vtd treat iova as gpa since dmar has been disabled during
> > > reset. (E.g In my test with Linux driver, desc map succeed but used fail.)
> > That would be a Linux bug, the devices below the IOMMU have to be reset
> > first.
> > 
> > Paolo
> 
> Maybe, but I meet the same issue if I do a "system_reset". I think we should
> guarantee IOMMU was reset after any pci devices in this case too?
> 
> Thanks
> 

system reset would indeed appear to be a big issue.

-- 
MST