Re: [PATCH 00/34] AMD IOMMU driver

[Joerg Roedel <joerg.roedel@amd.com>]

On Fri, Jun 27, 2008 at 10:18:08AM +0200, Ingo Molnar wrote:
> 
> * Joerg Roedel <joerg.roedel@amd.com> wrote:
> 
> > Hi,
> > 
> > this is the first post of the initial driver for AMD IOMMU hardware. 
> > The driver is tested on hardware with various loads (disk and network) 
> > and showed no problems.
> > 
> > It currently supports DMA remapping using the dma_ops API in the x86 
> > architecture code. It also supports isolation of device DMA address 
> > spaces as far as the hardware allows that (means each device can get 
> > its own DMA address space and can't access the DMA memory of other 
> > devices).
> 
> the code looks very clean. I guess down the line we might want to think 
> about making all the externally visible knobs transparent: there should 
> be only one way to disable the iommu via the boot command line (be that 
> an AMD or Intel one), etc.
> 
> i have created a new -tip topic for this: tip/x86/amd-iommu and have 
> applied your patches there - thanks Joerg. You can pick up the 
> integrated tree via:
> 
>   http://people.redhat.com/mingo/tip.git/README
> 
> please check whether the integrated end result still works fine.
> [ Obviously nobody but you can test this on real hw ;-) ]

Ok, I am testing the tip-branch for AMD IOMMU currently on real hardware
and everything looks very good. I will continue with load tests over the
weekend to be sure.

Joerg

-- 
           |           AMD Saxony Limited Liability Company & Co. KG
 Operating |         Wilschdorfer Landstr. 101, 01109 Dresden, Germany
 System    |                  Register Court Dresden: HRA 4896
 Research  |              General Partner authorized to represent:
 Center    |             AMD Saxony LLC (Wilmington, Delaware, US)
           | General Manager of AMD Saxony LLC: Dr. Hans-R. Deppe, Thomas McCoy