Re: [PATCH] iommu/amd: Fix unity mapping initialization race

[Wan Zongshun]

On 2016å07æ11æ 15:19, Joerg Roedel wrote:
> On Sun, Jul 10, 2016 at 07:40:53PM +0800, Wan Zongshun wrote:
> > Do you mean we need enable the V and TV bits to DTE entry after all
> > DTEs tables were initialized completely?
>
> Yes, this is what my patch does and what fixes the bug that was
> reported on machines which have unity-mapping entries.

Okay, this patch should also better to general case not only unity-mapping.

How about the interrupt remap function? Do we need same considering for 
IV bit enable for interrupt remap?

> > I checked this function 'init_device_table_dma', and find it just set
> > V and TV bit, to set translation info valid and DTE bits127:1 valid.
>
> Right, if no other bits are set this blocks all DMA from the gives
> device-id.

Sorry, why you still say this 'init_device_table_dma' can block DMA?
I just think this function will enable DMA transfer, since  we set the V 
and TV bits, right? or I misunderstand what "block DMA" mean?

> > So I just think all things it should to do are to allow DMA access,
> > GPA-to-SPA translation should be active, why you add function
> > comments below is to not allow DMA access and suppress all page
> > faults?
> >
> > /*
> >   * Init the device table to not allow DMA access for devices and
> >   * suppress all page faults
> >   */
>
> Yeah, that comment needs to be updated. Not all DMA is blocked and
> page-faults are not suppressed at all. Thanks for noticing.
>
>
>
> 	Joerg