On 11/07/12 10:04, Benjamin Herrenschmidt wrote:
> On Wed, 2012-07-11 at 09:55 +1000, Alexey Kardashevskiy wrote:
>> On 11/07/12 08:26, Scott Wood wrote:
>>> On 07/10/2012 12:51 AM, Alexey Kardashevskiy wrote:
>>>> The patch enables VFIO on POWER.
>>>>
>>>> It literally does the following:
>>>>
>>>> 1. POWERPC IOMMU support (the kernel counterpart is required)
>>> [snip]
>>>> +/* -------- API for POWERPC IOMMU -------- */
>>>> +
>>>> +#define POWERPC_IOMMU 2
>>>> +
>>>> +struct tce_iommu_info {
>>>> + __u32 argsz;
>>>> + __u32 dma32_window_start;
>>>> + __u32 dma32_window_size;
>>>> +};
>>>> +
>>>> +#define POWERPC_IOMMU_GET_INFO _IO(VFIO_TYPE, VFIO_BASE + 12)
>>>
>>> Is there a more specific name that could be used for this? Not all
>>> PowerPC chips have the same kind of IOMMU.
>
>>
>> Ben, is is SPAPR? BOOK3S?
>
> So we have varieties of different iommus in the kernel indeed, we
> probably want the info ioctl to reflect that. I would call this one
> spapr_tce.
VFIO provides such ioctl actually and it is it who returns POWERPC_IOMMU as an
IOMMU type.
ok. SPAPR_TCE.
> Also we will want a few other things, dunno if that's reflected here,
> or whether the ioctl is easily extendable, but in the long run we will
> need:
>
> - Ways to tell KVM about association between a liobn (logical bus
> number as used in H_PUT_TCE) and an iommu so we can implement the real
> mode H_PUT_TCE properly.
>
> - We will need some conduit to implement the "DDW" APIs (part of PAPR
> allowing the guest to control the DMA windows, ie, by creating new
> windows in 64-bit DMA space, with different page sizes etc....). So you
> may want to make it clear that the above provides information about the
> "base window" specifically.
So the current one would be SPAPR_TCE_32?
--
Alexey
