Re:RE: [PATCH 1/3] KVM: pci passthrough un-page-aligned mmio device on machines w/o VT-d

Re:RE: [PATCH 1/3] KVM: pci passthrough un-page-aligned mmio device on machines w/o VT-d

[刘晓建]

2008-12-16，"Han, Weidong" <weidong.han@intel.com> wrote：
>Pls note that multiple devices may share one page for their MMIO, and they may
> be
> assigned to different guests, there is potential isolation issue. So we don't
> allow to assign these devices.

In my view, we can modify the get_real_device() to open the file /proc/iomem to
find out whether the "potential issue" is a real one, instead of disabling this
feature all the times.

Xiaojian Liu

RE: Re:RE: [PATCH 1/3] KVM: pci passthrough un-page-aligned mmio device on machines w/o VT-d

[Han, Weidong]

刘晓建 wrote:
> 2008-12-16，"Han, Weidong" <weidong.han@intel.com> wrote：
>> Pls note that multiple devices may share one page for their MMIO,
>> and they may be assigned to different guests, there is potential
>> isolation issue. So we don't allow to assign these devices.
> 
> In my view, we can modify the get_real_device() to open the file
> /proc/iomem to find out whether the "potential issue" is a real one,
> instead of disabling this feature all the times.
> 
> Xiaojian Liu

As I know, there are few devices that have unpage-aligned mmio, and they are almost legacy. What's more, it's a little weird to map unpage-aligned memory slot, though I don't know what's problem on it. In short, I don't perfer to assign these devices.

Regards,
Weidong

Re:RE: [PATCH 1/3] KVM: pci passthrough un-page-aligned mmio device on machines w/o VT-d

[刘晓建]

> In my view, we can modify the get_real_device() to open the file /dev/iomem to
> find out whether the "potential issue" is a real one, instead of disabling this
> feature all the times.

Sorry, I made a mistake. I mean the file /proc/iomem, not /dev/iomem.

Regards,
Xiaojian Liu